Hardcore Software by Steven Sinofsky (Audio Edition)

The planning for Windows 8 was moving right along. But something wasn’t right as we wrapped up Windows 7 activities at CES 2010. It was looking more and more like the plans and the way the ecosystem might rally around them would yield a watered-down result—it would be Windows and a bunch of features, or perhaps irreconcilable bloat. The way the ecosystem responded to touch support in Windows 7 concerned me. How do we avoid the risk of a plan that did too much yet not enough? Oh, and Apple scheduled a “Special Event” for January 27, 2010, just weeks after a concerning CES.

Back to 097. A Plan for a Changing World [Ch. XIV]

In early January 2010, I was walking around the show floor at CES the evening before opening day as I had routinely done over the years. CES 2010 was a mad rush to build a giant city of 2,500 booths only to be torn down in 4 days. This walk-through gave me a good feel for the booths and placement of demonstrations. It was just two months after the launch of Windows 7. Walking around I made a list of the key OEM booths to scope out first thing in the morning. It wasn’t scientific but visiting a booth had always been an interesting barometer and sanity check for me compared to in-person executive briefings. Later in the week I would systematically walk most of the show and write a detailed report. The next morning, along with the giant crowds, I made my way through a few dozen booths with the latest Designed for Windows 7 PCs mixed in among the onslaught of 3D-television controlled by waving your hands which garnered much of the show’s buzz.

The introduction of touch screens was a major push for Windows 7 and there was genuine excitement among the OEMs to offer touch as an option, though few, okay none, thought it would be a broadly accepted choice. Touch added significantly to the price. With two relatively small suppliers for the hardware, OEMs were not anxious to make a bet across their product line. TReller, the new CMO and CFO for Windows, made a good decision to provide strategic capital to one component maker to ensure Dell would make such a bet.

There were touch models from most every PC maker, but they were expensive—most were more than $2,500 when the typical laptop was sub-$1,000. For the OEMs this was by design. If a buyer wanted touch there was an opportunity to sell a high-end, high-margin, low-volume device. OEMs had been telling us for months that this was going to be the case, but it was still disappointing.

Taking advantage of Windows 7, and wholeheartedly, were a sea of Windows 7 “slates” all based on the same design from the combination of Intel and Pegatron, an ODM. These slates were essentially netbooks without keyboards. They fit the new Intel definition previously described—MID or mobile internet device. They were theoretically built as consumption-oriented companions to a PC. They were shown reading online books, listening to music, and watching movies, though not particularly high resolution or streaming given the meager hardware capabilities. All of them were relatively small and low-resolution screens. To further emphasize the Intel perspective, they also launched AppUp for Windows XP and Windows 7, a developer program and early content store designed to support rights management and in-app purchase as one might use for books and games.

The buzziest slate was from Lenovo, not a product announcement, but a prototype model kept behind glass at the private Lenovo booth located in a Venetian Hotel restaurant. The “hybrid slate” Lenovo U1 was a 11” notebook that could also be separated from the keyboard and used as both a laptop and a slate. As a laptop, the Windows 7 PC used a low-power Intel chipset, a notch slightly above a netbook. Detached as a slate, the device ran a custom operating system they named Lenovo Skylight based on Linux running on a Qualcomm Snapdragon chipset. The combination weighed almost 4lbs. The Linux tablet separated from the Windows-based keyboard PC somewhat like the saucer section of the Starship Enterprise separated from the main ship. Lenovo built software to sync some small amount of activity between the two built-in computers, such as synchronizing bookmarks and some files. Economically, two complete computers would not be the ideal way to go, bringing the cost to $1000. Strategically for Microsoft this was irritating.

Nvidia, primarily known for its graphics cards used by gamers, was always an interesting booth. Nvidia was really struggling through the recession and would finish 2010 with revenue of $3.3 billion and a loss of $60 million. As it would turn out it was also a transformative year for the company, and for one of the most legendary founders in all of the PC era, Jensen Huang. To put Nvidia in context, my very first meeting with Intel when I joined the Windows team in 2006 was about graphics, because of the Vista Capable fiasco. Intel was digging their heels in favoring integrated graphics and was not at all worried about how their capabilities were so far behind what Nvidia and ATI, another graphics card maker, were delivering, which Intel viewed as mostly about games. The rub was that AMD, Intel’s archrival, had just acquired ATI for $5.4 billion making a huge bet on discrete (non-integrated) graphics, which was what Nvidia focused on. Intel seemed to believe that the whole issue of graphics would go away as OEMs would simply accept inferior graphics from Intel because it was cheaper and easier while Intel improved integrated graphics over time, albeit slowly. A classic bundling strategy of “a tie is a win” that would turn out to be fatal for Intel and an enormous opportunity for ATI/AMD, Nvidia, and Apple. Intel could have acquired Nvidia at the time for perhaps $10-12B if that was at all possible.

In their booth Nvidia was showing off how they could add their graphics capabilities to the Intel ATOM processor and dramatically speed it up. Recall, ATOM chips struggled to even run full screen video at netbook screen resolutions. With Nvidia ION it was possible to run flawless HD video. Nvidia made this all clear in a “Netbook Nutrition Facts” label that they affixed to netbooks running Nvidia graphics. This was a shot across the bow at Intel, but Nvidia had an uphill battle to unseat bundled graphics from Intel. Nevertheless, we were acutely aware of the strong technical merits of Nvidia’s approach, Jensen’s incredible drive, and the needs of Windows customers. This will prove critically important in the next chapter as we further work on non-Intel processors.

More disappointing was how the OEMs chose to demonstrate touch capabilities in Windows 7. We provided OEMs with a suite of touch-centric applications, such as those demonstrated a year earlier—mapping, games, screen savers, and drawing, for example. We even named it Windows Touch Pack. The OEMs wanted to differentiate their touch PCs with different software and viewed the Touch Pack we provided as lame. Touch had become wildly popular in such a short time because of the iPhone, and the iPhone was a consumer device used for social interaction, consumption of media, and games.

With that frame of reference, the OEMs wanted to show scenarios that were more like an iPhone. The OEMs were going to do what they did, which was to create unique software to show off their PCs. This is just what many in the industry called crapware or, as Walt Mossberg coined in his column, craplets. To the OEMs this was a value add and important differentiation. Worst of all, there was no interest from independent software makers who had dedicated few, if any, resources to Win32 product updates, let alone updates specific to touch that was exclusive to the new Windows 7. We never intended to support Windows 7 touch on earlier versions of Windows—something developers ask of every new Windows API. This compatibility technique we often relied on but was a key contributor to Windows PC fragility and flakiness, also referred to colloquially as DLL Hell as previously described.

The overall result: Windows 7 provided no impetus for third party-developers, and we failed to muster meaningful third-party support for any new features, including touch.

What I saw was a series of new OEM apps taking a common approach. These apps could be called shells in typical Microsoft vernacular, an app for launching other apps—the Start menu, taskbar, and file explorer in Windows constitute the shell. In this case, these new shells were usually full-screen apps that had big buttons across the top to launch different programs: Browsing, Video, Music, Photos, and YouTube. These touch-friendly shells did not do much other than launch a program for each scenario, the browser, or simply a file explorer. For example, if Music was chosen then a music player, created by the OEM or chosen because of a payment for placement, would launch to play music files stored on or downloaded to a PC. The use of large touch buttons was thought to give the PC user interface a consumer feel. The browser used part of Internet Explorer but not the whole thing, just the HTML rendering component; for example, it did not include Favorites or Bookmarks usually considered part of a browser. This software, as well intentioned as it was, would fall in the crapware category. This was expected, but still disappointing.

In booths with mainstream laptops we were offered a glimpse into the changing customer views of what made a good laptop. The Windows 7 launch was just a short time ago, so most laptops had incremental updates, primarily with the new version of Windows and perhaps a slight bump in specs. Some show attendees picked up the laptops and grimaced at the weight and thickness—these machines were hardly slick. It had been two full years since the introduction of the MacBook Air and the PC industry still did not have a mainstream Windows PC that fit in the famous yellow envelope wielded on stage by Steve Jobs. The prevalence of Wi-Fi in hotels and workplaces had changed the view that a work PC needed all the legacy ports present on Windows computers, and the fear of dongles had faded. Most every Windows PC still shipped with an optical drive, significantly increasing the height of the PC. Show-goers were openly commenting on the size, weight, and lack of portability of Windows PCs compared to the MacBook Air, which was what many were fully lusting for and based on the numbers had already switched to. For the two years since the MacBook Air launch the PC makers were pre-occupied with netbooks.

Apple broadened their product line with an even smaller and lighter MacBook Air using an 11.6” screen. They also added MacBook Pro models that were more in line with Windows PCs when it came to hardware, for example they included optical drives and discrete graphics. These laptops were expensive relative to mainstream business PCs running Windows. For example, the Dell XPS 15 debuted in 2010 was praised by the press, though only relative to other PCs. Compared to MacBooks, the Dell lost out for a variety of reasons including noise, keyboard feel, trackpad reliability, screen specs, and more. In reviewing the new Dell XPS, AnandTech, a highly regarded tech blog, said:

We've lamented the state of Windows laptops on numerous occasions; the formula is "tried and true", but that doesn't mean we like it…what we're left with is a matter of finding out who if anyone can make something that truly stands out from the crowd.

Of course, if we're talking about standing out from the crowd, one name almost immediately comes to mind: Apple. Love 'em or hate 'em, Apple has definitely put more time and energy into creating a compelling mobile experience.

If only there had been a Windows PC like MacBook Air for Windows 7. The closest we had was the low-volume and premium priced Sony VAIO Z that was new for 2010. This model featured a solid-state drive like the MacBook Air but was much heavier and larger than the famous envelope and could easily top $3000 when fully specified.

This knocked the wind out of me. I was happy that there were Windows 7 PCs at the show. But seeing the reaction to them only reinforced my feeling that the ecosystem was not well. I was hardly the first Windows executive to bemoan the lack of good hardware from partners, especially hardware that competed with Apple. Six years earlier, before the MacBook Air but also around CES, my predecessor JimAll sent a polemic, intentionally so, to BillG, Losing our way…in which he said, “I would buy a Mac today if I was not working at Microsoft.” As with many of these candid emails, this one made its way to an exhibit in a trial—it added nothing to the case, but such salacious mails that have little by way of legal implications are often used to attempt to gain leverage for settlements or stir emotions in a courtroom.

By the time the show ended, I wrote up my CES 2010 trip report as I had been doing since about 1992 or so. I loved (and still love) writing up these reports. No matter how down I was or how boring I thought the show was, I amped up the excitement as though it was the first CES I ever attended. I think it is important to do because otherwise the aging cynicism that seeps in—often seen in the press covering the event—makes it too easy to miss what might be a trend. My report ran 20 pages with photos and covered everything from PCs of all shapes and sizes to 3D television to iPhone accessories (lots of those.)

Not lost on everyone at the show was how Apple loomed over the show even though it had zero official presence. The vast majority of accessories located the edges of the Sands Exhibition Center were cases, docks, cables, and chargers for the iPhone. The camera area was filled with accessories to turn an iPhone into a production camera, such as a Steadicam for an iPhone that I even mocked in my report, oops that was way off base. The real news, however, was that the gossip of a forthcoming Apple tablet was everywhere. Nearly every news story about the show mentioned the one tablet that wasn’t there.

The week before CES in a classically Apple move, invites were sent to the press for an Apple Special Event to be held January 27, 2010. Leading up to the event, rumors swirled around how Apple would introduce a new touch computer and eventually converged on a 10” tablet to be based on iPhone OS. As we see today, the rumors were wildly off base, and only days before did the rumors mostly match the eventual reality.

After the first day of CES and our contribution to the main keynote by SteveB, that evening, I was in a mood and not a good one. I skipped yoga (at the awesome Vegas HOT! studio) and a previously planned celebratory team dinner that I arranged (!) to write a memo. Maybe memo isn’t the right word. It was another dense thought-piece from me, a 6000-word SteveSi special for sure. I think JulieLar, MikeAng, and ABurrows are still angry with me for missing that dinner.

I knew mainstream PCs were selling well. PCs from Dell, Acer, and HP met the sweet spot of the market, even among a marketplace overdosing on netbooks. People were over the moon for Windows 7. Office 2007 was doing quite well, despite not having compatibility mode, which had receded to a non-issue. Windows Server and associated products were going strong even though cloud computing had taken over Silicon Valley. Even Bing was showing signs of life six months or so into its rebranded journey, which had been managed by Satya Nadella (SatyaN) for the past few months. These all made for the bull case that Microsoft was making progress.

For customers and the tech industry, well, their attention had by and large moved to non-PC devices, Apple versus Google in phones, and the potential new tablet form factor. Microsoft launched the HTC HD2 running Windows Mobile 6.5 at CES. It was a well-received and valiant effort while Windows Phone 7 progress continued even in a decidedly iPhone world.

I remained paranoid, Microsoft paranoid. The dearth of premium PCs to compete with Apple and the lackluster success with touch gave me pause. Even though we had tried to right all the wrongs of introducing new hardware capabilities with new approaches, we had failed. There were no Windows 7 touch apps. There never would be. It was totally unclear if the industry would ever come together to create a MacBook Air, and even if it did it was unlikely to have the low cost, battery life, and quality over time of a Mac, even though it too was running on Intel. The browser platform consumed most all developer attention, with iPhone and then Android drawing developers from the browser. Apple’s App Store, 18 months old, had 140,000 apps up from 500 at launch, growing at a rate of 10 to 20 percent per month. To put that in perspective, during the Windows 7 beta I shared what I thought was an astoundingly large number—883,612 unique Windows applications seen across the massive beta test. By mid-2022, the Apple store had more than 3.4 million apps!

A sense of dread and worry came over me when I thought about our converging Windows 8 plans. It wasn’t a panic attack. It was more me wondering if I failed to give clear enough direction to the team about how big a bet we were willing to make on Windows 8. Was I too subtle, which I was definitely known to be, and gave too much room for the plan to turn into an “and” plan? An “and” plan means we would do everything we would have normally done, “and” also take into account all the new scenarios. Such plans are easy to sign up for because they don’t involve tradeoffs. At the same time, they force no decisions and lacked the constraints that yield a breakthrough design.

An “and” plan is what I saw across the CES show floor and even from Intel. It is a plan where the PC is a Windows PC and the new stuff, as though the new stuff is just another app one could take or leave. The extra launcher shells, content stores, and touch interface were bolted on the side of Windows and did not substantially alter the value proposition that was under so much pressure. Related, an “and” plan would also presume a traditional laptop or desktop PC as the primary tool, which itself would prove as limiting down the road as it was right then.

We needed a plan with a point of view about how computing would evolve. That point of view needed to be built upon the assumption that Win32 was not the future, or even the present. We needed to realize that an app and content store were of paramount importance. We needed to account for shifting customer needs in their computing device. We needed a plan that assumed the web is the defining force that changed computing and smartphones were themselves changing the web. We needed a plan that substantially changed the futzing, viruses, malware, and overall fragility of the PC so it acted more like a consumer electronics device. Most of all we needed a plan to engage developers in a new way to become interested Microsoft’s solution this opportunity, and not the solutions from Apple or Google.

We had the ingredients of a plan that might be adequate to quell the forces of disruption. Would executing such a plan be possible within the context of Microsoft? What does a plan look like when it also involves Windows itself?

The Windows 8 framing memo I wrote and the planning memo and process JulieLar had started contained the core of a point of view. We had set out to investigate “alternative hardware platforms,” potentially “distributing applications in a store,” designing “a modern touch experience,” and much more. As interesting and innovative and outside the box as these were, they all suffered the same risk. The problem was that for the team they were viewed as additions to Windows, as add-ons to the next release, as nice to have, an addition to. We had—and this is where the discomfort came from—created a classic strategy of the incumbent when faced with disruption. Our strategy treated the forces of disruption—in this case mobile hardware platforms, modern user interface, and the app ecosystem—as incremental innovations in addition to what we had. We were thinking that we would be competitive if we had Windows and other features.

This is not how disruption works.

Disruption, even as Christensen had outlined a decade earlier, is when the new products are difficult for the incumbent to comprehend and are a combination of inferior, less expensive, less featured, and less capable, but simultaneously viewed by customers as superior replacements.

As a team, we had faced this before, and I had spent the better part of three releases of Office fending off those claiming Office was being disrupted by inferior products built on Java, components, or browsers. In all those instances, I stuck to what I believed to be the case and defended, effectively, the status quo until we took the dual risks of expanding Office to rely on servers and services and then redefining the user experience for the product. Patience paid off then. The new technology threats were immature at best and a distraction at worst. 

What gave me the confidence to believe, as they say, this time was different? Was it my relative inexperience in Windows of just one release? Naivety? Arrogance? Was it envy of Apple or Steve Jobs?

It was none of those things.

It was simply the combination of the shortcomings I was seeing on the CES show floor and the fact that we’d all been using iPhones. We were not talking about a theoretical disruption where someday Java Office could have all the features people wanted in Office or that someday all the performance and UI issues would be squeezed out of the browser. The disruption had already happened, but the awareness of that was not shared equally by everyone involved.

To use a phrase attributed to William Gibson, “The future has arrived—it’s just not evenly distributed yet.”

Windows had 100% share of the Windows PC market and 92% share of the PC market. iPhone and Android (phones, and soon other form factors) were making a compelling case that the Windows position, as a percentage of computing, was declining. This decline in computing share would only accelerate.

In technology, if you are not growing then you are shrinking.

Holed up in my hotel room while the rest of the team had dinner, I banged away at my netbook keyboard a screed about the future of the web, apps, and how consumers would interact with content—and, importantly, ways in which the PC was deficient. I wrote 6,000-plus words, making the case that the iPhone was so good the internet was going to wrap itself around the phone rather than the phone becoming a browser for the existing internet—the web would tailor itself to the phone via apps. The experience of the web was not going to be like it previously had been—the iPhone was not going to deal with the desktop web and attempt to squeeze it on to the phone. The browser itself would become a form of legacy, “the 3270 terminal” of the internet era. The 3270 is the IBM model number for a mainframe terminal, replaced by PCs and applications.

As the office workers I gave first PCs to in 1985 did not realize, the business forms they filled out with their Selectric typewriters would change to work on a PC, not that the PC would get good at filling out carbon paper forms. A key part of a technology paradigm shift is that the shift is so significant that work changes to fit within the new tools and paradigm. There’s only a short time when people clamor for the new technology to feel and behave like the old.

I sent the memo to a few people that night. It received a bit of pushback because it felt like what we were already doing, only restated. I kept thinking this was how disruption really happened—teams went into denial, say “We got this” and believe adding a bit more would make the problem (or me) go away. IBM added a 3270 terminal emulator to the original PC for this reason, but that didn’t make people want a PC more or even use it; rather, they complained that the terminal seemed more difficult to use than a standard 3270 terminal. The PC I most frequently installed in the summer of 1985 was the IBM PC XT/3270, a hybrid mess if there ever was one.

It was clear we were falling into the trap of thinking incrementally when the world around us was not only changing but converging. The Windows team could no longer think of modern mobile platforms as distinct from the PC. People’s technology needs were being met by phones and were moving away from the PC.

This was not a call to turn the Windows PC into a phone. Rather this was a chance to up-level the PC, to embrace the web, embrace the app store model, embrace the shifting hardware platform and associated operating system changes, and embrace internet technologies for application development.

Most of all it was about “apps” except I couldn’t quite find the right word to use. Apple talking about “apps” for the past couple of years really bugged me. When I was in Office, we were the Applications (or Apps) group at Microsoft and an app was just the techie word for another more broadly used techie word, program. Word and Excel were apps and I was trying to define a new kind of app. App was such a weird word that even our own communications team didn’t think we should use it. Lacking a better word that night in the hotel, I called the memo “Stitching the Tailored Web” which was a horrible name and worse metaphor that took a lot of time for me to explain to people in person.

The key ingredients of this new world as I described it were a computing device, content store, and the tailored web itself. The computing device encompassed an interaction model for the device and for the new tailored apps. The content store is where users obtained apps created and distributed by developers along with consumable content such as movies, books, and music. The tailored web included the APIs used by developers and the OS platform, the new tools required by developers, as well as the core technologies of Internet Explorer used to create these applications and to browse the traditional web. All three came together to create what I described as the more “consumable internet.” I even had a fancy diagram I made with PowerPoint to illustrate the point.

My main point was that mobile phones were so huge and such a force that even the internet would change because of phones, the iPhone in particular. The notion of going to web sites would be viewed as quaint or legacy compared to apps. Today, there are plenty of people who do not take this point of view to be where we landed. There are those that use desktop PCs and live in a desktop browser and defiantly believe that to be the only way to get real work done. They use the highest-end PC tools such as Visual Studio, Excel, AutoCAD, and Adobe Premier and did not (and still do not) see phones or mobile software replacing those any time soon. Yet there is no disputing the vastness of mobile. Depending on the source (and country) about three-fourths of all web traffic is mobile-based, and about 90% of time spent on a mobile device is in an app. With this dramatic change in how the web was used, it is fair to say that the web became a component of mobile, not the other way around.  The browser-based Web 2.0 juggernaut would be subsumed by the smartphone.

The most dramatic expression of this would not happen for another 2 years. In the summer of 2012, Facebook, the posterchild of Web 2.0, dramatically pivoted from desktop browsing as the primary focus for the Facebook experience to the Facebook mobile app. That legendary pivot is often touted as both remarkable and responsible for the success and reach the company subsequently achieved.

The Tailored Web memo was much better delivered as a passionate call to the team. I waited until just after Apple’s scheduled event and later that same day held a meeting with our senior managers (the 100 or so group managers across all of Windows, each the leader of dev, test, pm for a feature team.) This was my favorite meeting and one I held at least quarterly going back to the late 1990s in Office.

Working with Julie, this became our own version of a pivot. It was not as much a pivot as simply a focusing effort. We still had about three months until the vision needed to be complete. This was by no means a major change. Rather it was relaxing the constraint that Windows 8 had to do two things, an “and” project, and it could focus on making sure we were set up for the future. The call to action was a call to prioritize the new over the legacy and interoperability between the legacy and new was not a priority. We wanted to break away from what was holding us back, a legacy neither of our primary competitors dealt with.

To the team, I wanted to make the case about apps by showing a series of screen shots of web sites versus Apps on various platforms and rhetorically asked “Which would you rather use?” Facebook on iPhone or a desktop? Crowded, ad-filled Bing on a desktop or the Bing search app on Windows Mobile? Outlook or the Mail app on iPhone? Reading a book in a browser or on a Kindle? A traffic map on a desktop browser or in the Microsoft Research SmartFlow traffic app?

The answer was obvious for all of us and remains obvious even today. The call to action was a doubling down on the developing planning themes “Defining a Modern PC Experience” and “Blending the best of the Web and Rich Client” with an emphasis on consumption and the internet. Specifically, it was “Windows, in all form factors, needs to be the best platform and experience to consume internet content while enabling ISVs and IHVs to build the software and hardware to do that.”

It was early in the potential for this disruption, and it was easy to point to thousands of tasks, features, scenarios, partners, business models, ecosystems, and more that required a PC as we knew it. Every situation is different, but it is that sort of defensiveness that prevented companies like Kodak, Blackberry, Blockbuster, and more from transitioning from one era to another.

It is easy to cycle from anxiety to calm in times like this. It was easy for me to look around and find affirmation of the “and” path we were on. Every stakeholder in the PC ecosystem would be far happier with incremental improvements to the status quo rather than embarking on huge changes with little known upside and significant immediate cost and potential downside. It would have been easy to fall back on the financial success we had and leave dealing with a technology shift for the future when technology specifics were clearer. It would, however, be too late as we now know. We would also lose the opportunity to lead a technology change that had just begun.

Worrying seemed abstract. Developing a clear point of view and executing on that seemed far more concrete and actionable.

The announcement from Apple and their tablet turned out to be far more than our poorly placed sources had led us to believe. What was it like in the halls of the Windows team on that day, January 27, 2010? It was not magical for us.

On to 099. The Magical iPad

Welcome to Chapter XIV. This is the first of two chapters and about a dozen remaining posts that cover the context, development, and release of Windows 8. Many reading this will bring their own vivid recollections and perspectives to this “memorable” product cycle. As with the previous 13 chapters and 96 posts about 9 major multi-year projects, my goal remains to share the story as I experienced it. I suspect with this product there will be even more debate in comments and on twitter about the experiences with Windows 8. I look forward to that. This chapter is the work and context leading up to the plan. Even the planning process was exciting.

Back to 096. Ultraseven (Launching Windows 7)

In the summer of 2012, I was sitting across from BillG at the tiny table in the anteroom of his private office on the water in Kirkland. The sun was beaming into my eyes. In front of me was one of the first boxes of Microsoft Surface RT, the first end-to-end personal computer, general-purpose operating system, and set of applications and services designed, engineered, and built by Microsoft. In that box sat the culmination of work that had begun in 2009—three years of sweat and angst. After opening it and demonstrating, I looked at him and said with the deepest sincerity that this was the greatest effort and most amazing accomplishment Microsoft had ever pulled off.

Later that same week, I had a chance to visit with Microsoft’s co-founder, Paul Allen (PaulA), at his offices at the Vulcan Technology headquarters across from what was then Safeco field. I showed him Surface RT. I previously showed him Windows 8 running on desktop using an external touch monitor. At that 2011 meeting he gave me a copy of his book Idea Man: A Memoir by the Cofounder of Microsoft and signed it. Paul was always the more hardware savvy co-founder, having championed the first mouse and Z80 softcard, and had been pushing me the whole release of Windows 8 on how difficult it would be to get performance using an ARM chipset and on the challenges of hitting a low-cost price point. Years earlier, Vulcan built a remarkably fun PC called FlipStart, which was a full PC the size of a paperback novel. Surface RT with its estimated price of $499, $599 with a keyboard cover, and a fast and fluid experience, resulted in the meeting ending on a high note. I cherished those meetings with Paul. I also shared with Paul, proudly, my view of just how much we should all value the amazing work of the team.

What I showed them both was the biggest of all bets. While not “stick a fork in it” done, by mid-2012 Microsoft seemed to have missed the mobile revolution that it was among the first to enter 15 years ago. In many ways Surface RT set out to make a new kind of bet for Microsoft—a fresh look at the assumptions that by all accounts were directly responsible for the success of the company. Rethinking each of those pillars—compatibility, partnerships, first-party hardware, client-computing, Windows user-interface, and even Intel—would make this bet far bigger and more uncomfortable than even betting the company on the graphical interface in the early 1980s. Why? Because now Microsoft had everything to lose, even though it also had much to gain.

With Windows 7, we knew we had a traditional release of Windows that could easily thrive through the full 10-year support lifecycle as we had seen with Windows XP. Windows 7 would offer a way to sustain the platform as it continued to decline in relevance to developers and consumers, while extracting value from business customers with little incentive to change.

Microsoft needed a new platform and a new business model for PC makers, developers, and consumers. The only rub? Any solution we might propose wasn’t something we could A/B test or release pieces at a time experimentally. Windows was the standard and wildly successful. It wasn’t something to experiment on.

While the company was 100 percent (or more) focused on Windows 7, we had started (drumroll for the codename please) Windows 8 planning five months earlier. We had the next step, a framing memo from me first. Then we had a planning memo from JulieLar, ready to go as soon as we all caught our collective breath after the release—no real time to celebrate more, and certainly no downtime.

Technology disruption is often thought of at a high level along a single dimension, but it is far more complex. Consider Kodak’s encounter with digital photography, Blockbuster’s battle with DVD-by-mail, or the news business’s struggle with the web. Great memes, sure, but one layer down each is a story of a company facing challenges in every attribute of their business, and that is what is interesting and so challenging.

Digital photos were more convenient for consumers, that was true. But also, the whole of Kodak was based on a virtuous cycle of innovation developed by chemical and mechanical engineers, products sold through a tightly controlled channel, and an experience relying on a 100-year-old tradition of memorialized births, graduations, weddings, and more. At each step of technology change to digital images, another major pillar of Kodak was transformed, sliced, and diced in a way Kodak could not respond. The magical machinery of Kodak was stuck because not one part of its system was strong enough to provide a foundation. Kodak didn’t need to also enter the digital market. The only market that would come to exist was digital. The only thing that made it even more difficult was that it would take a decade or more to materialize as a problem, and that during that time, many said, “Don’t worry. Kodak has time.”

While figuring out what to do next for Windows, we saw Blackberry facing a “Kodak moment.” Blackberry was not just a smartphone, but a stack of innovations around radios, a software operating system optimized for a network designed for small amounts of data consuming little power, a business model tuned to ceding control to carriers, and a keyboard loved by so many. In 2009, Blackberry still commanded almost 45 percent of the smartphone market, even though the iPhone had been out since 2007. That led many to find a false comfort in the near term and to conclude Blackberry would continue to dominate. Apple’s iPhone delivered a product that touched every pillar of Blackberry, not only Blackberry the product but Blackberry the company. Sure, Apple also had radio engineers, but they also had computer scientists. Sure, Apple also worked with AT&T, but Apple was in control of the device. And Apple was, at its heart, an operating system company. Blackberry had some similar elements up and down the entire product stack, but it wasn’t competitive. At its heart, Blackberry was a radio company. Blackberry seemed to have momentum, but market share was declining by almost 1 percentage point per month.

The smartphone, iPhone and Android in particular, was disrupting Windows. Though, not everyone thought that to be the case. Some said that phones did not support “real work” or “quality games.” The biggest risk to a company facing disruption is to attempt to dissect disruptive forces and manage each one—like add touch or apps to a Blackberry with a keyboard. The different assumptions and approaches new companies take only strengthen over time, even if eventually they take on characteristics of what they supplant. The iPhone might never be as good as the PC at running popular PC games, but that also probably won’t matter.

I’d lived through graphical operating systems winning over character mode, PC servers taking over workloads once thought only mainframes could handle, and now I found myself facing the reality that browsers could do the work of Windows, relegating Windows to a place to launch a browser and not even Microsoft’s.

Every aspect of the Windows business faced structural challenges brought on by smartphones. Compounding the challenge, Apple was competing from above with luxurious and premium products vertically integrated from hardware to software. Google, with Android and Chrome, was competing from below with free software to a new generation of device makers. The PC and the struggling Windows phone were caught in the middle, powerless to muster premium PCs and unable to compete with a free open-source operating system.

The browser had already shifted all new enterprise software development away from the hard-fought victory of client-server computing. Anyone who previously thought of building a new “rich client” application with Visual Basic or some other tool was now years into the transition to “thin client” browser software. The Windows operating system was in no way competitive with smartphone operating systems. The way to develop and sell apps had been reinvented by Apple. The Win32 platform was a legacy platform by 2009, the only debate was how long that had been the case. A legacy platform does not mean zero activity, but it does mean declining and second or third-priority efforts.

The partnership between the Windows operating system and hardware builders had fundamental assumptions questioned by Android. OEMs supporting Android not only received the source code but were free to modify it and customize it to suit their business needs. That was exactly what Microsoft fought so hard against in the DOJ trial. The touch-based human interaction model developed by Apple and the large and elegant touchpad on Mac were far more approachable and usable than what was increasingly a clunky mouse and keyboard or poor trackpad on Windows. The expectations for a computer were reset—a computer should always be on, always connected, never break or even reboot, free from viruses and malware, have access to one-click apps, while easy to carry around without a second thought. That supercomputer in your pocket is connected to built-in operating system services providing storage, backup, privacy, security, and more. Like Kodak, many said to slow down, that change would not happen so quickly. Many said we could “add this” or “change that.” But it was not that simple, nor was it going to be that easy. The market was moving quickly.

The march to the next billion shifted, in the blink of an eye, to smartphones and soon non-Windows tablets, entirely skipping the PC. The entrenched PC customers were as excited as ever to upgrade to Windows 7 by buying a new PC or upgrading their existing PC. Everything new in Windows 7 was viewed through the lens of the past. How did we used to do that and did Windows 7 make it easier? Those skipping the PC didn’t concern themselves with improvements in a new version of Windows because they didn’t know about the old version. PC enthusiasts were still asking for features to handle managing files and keeping multiple PCs in sync while phone or browser users were seamlessly connected to a cloud and gave no thought at all to where files were stored. Buying a second iPhone or replacing a broken one proved astonishingly simple, because of the integration with services.

Microsoft’s answer to iPhone and Android 1.0 was still being developed. Windows Phone 7, originally codenamed Photon, would be released a year after Windows 7. The mobile team was heads down trying to get that release done, which had been in the works since Windows Mobile 6.0 shipped in early 2007. By the time Windows Phone 7 shipped, the Windows 8 team would have completed the first of three milestones. The next full chapter will detail the challenges this lack of synchronization created. We had Windows Phone 7 on a mission to get to market and compete and in doing so was pre-booking the next release; Windows 8 attempted to plan a major release that in theory would lend support to the mission of modernizing the codebase which was desperately needed by Windows Phone 7 competing with the modern Linux kernel and OS X.

Meanwhile, there were few new customers being brought to the cash-producing business of the Windows PC. There we were with the metaphorical Sword of Damocles hanging above us. The best days of the PC were behind us. It wasn’t just obvious technically—the value propositions of mobile devices were in no way compatible with the Win32 and PC architectures. The only question was when things would truly turn for the worst.

The risks were infinitely high. We could do something radically different, potentially choosing wrong and heading right into failure even accelerating what was inevitable. Or we could slow walk as long as possible, releasing incremental features for as long as possible hoping to find some future solution to our quagmire. Doing both is a classic upper management answer that with near certainty results in a combination of a bloated and confused near term that keeps trying to pull features from the new, and a brand new product that develops a major case of second-system syndrome. The latter is a classic defined in the standard developer book Mythical Man-Month, and what we experienced with numerous Microsoft projects from Cairo to Longhorn when the next generation attempts to solve all known problems at once.

It is why so many default to preserving the cash cow as the most logical strategy, maintaining the success already achieved. Usually this comes with raising prices along the way for existing customers, who are by and large captive. In the short term, things continue to look great, and a crisis will be averted while the crisis is left to a future team. In fact, in the short term our most ardent fans would continue to carry on as though everything is normal. They are part of the same bubble as us.

Either way, the world outside of the Windows PC shifted to mobile. We were at a fork in the road and had to pick a direction.

The thing about disruption when it is happening is that you alternate between over-confidence and paranoia. Financially, the Windows PC was secure. Emotionally, the Windows PC was fragile. From a product perspective, not only was there little left to do on the current path but doing anything would annoy both the disinterested customers and ardent fans more than it would encourage them or any new buyers. Over-confidence leads one to incremental plans, assuming the existing business is so strong as to not worry. Paranoia makes it difficult to identify the precise nature of disruption and to calibrate a response.

For all the insights that Innovator’s Dilemma captured, the one thing it did not seem to predict was that even after disruptions companies can continue to make vast sums of money for years to come from those disrupted products. Technologists tend to think disruption causes products to sort of go away, but as any private equity investor will tell you, that is not the case. That’s why a misplaced confidence in incrementalism almost always wins out over a bold, risky bet, at least for a while.

We were determined to find a way to do better than to slowly lose relevancy as we’d seen happen to IBM.

Planning Windows 8

We kicked off a planning process just as had been done for all the other releases described in this work—from Office 2000 all through Windows 7. By now this team had fifteen years of product planning and results that had transformed a set of bundled applications into a suite of interconnected products and then to a platform of products, servers, and services, and then reinvented the user experience setting these tools up for at least another decade. We then aimed this same process at the challenges of the Windows business, and by all accounts set that product up for another decade.

About three months before Windows 7 released to manufacturing, I sent out the first step in planning Windows 8, the obvious name. “Building on Windows 7” outlined the state of the business and product, suggested where we would invest for the next project, detailed competitive threats, and defined what success would look like.

As with all the other framing memos I wrote, this one began upbeat. How could it not? The release of Windows 7 was a super positive experience for the team, the company, and the broad ecosystem. The memo celebrated all we accomplished as a team and with Jon DeVaan’s leadership significantly improving our engineering capabilities and productivity. Promise and deliver.

Then the reality. The recession had slowed the growth of PC sales and analysts were taking forecasts down. The damage netbooks had done to pricing leverage was significant, resurrecting the Linux desktop and lengthening Windows XP in market. Netbooks also masked the secular weakness in PCs that was also seeing as lengthening of the PC refresh cycle in large companies. PCs were being tasked to last longer without loading a new version of Windows or Office. Emerging markets and software piracy were proving highly resistant to most every effort. The sabbatical I took living in China in 2004 taught me firsthand just how impossible it was going to be to thwart piracy of desktop Windows and Office, our primary sources of revenue. Windows Live was, as we say, making progress, but Google was making much more progress much faster, and our marquee products, Hotmail and Messenger, were losing share and fast. Releasing SkyDrive (now OneDrive) with excellent integration of the new Office Web Applications (Word, Excel, PowerPoint running in the browser with compatibility with desktop Office) was a significant bright spot, but it came with very high operational costs and immense pushback from Office that feared cannibalization. There were even changes in how PCs were being sold that were proving highly problematic. OEMs were increasingly relying on nagware or trial products, bloating the PC user interface, and putting a ton of pressure on the overall experience of Windows. The retail channel, struggling from the recession, was anxious because they had not yet mastered selling smart phones, the hot category, which increasingly looked like a carrier-specific play. This in turn caused retailers to become part of the equation of pre-loaded software, thus further eroding the experience.

Competitively, all I asked us to focus on was Apple and Google. Most would see Apple as a competitor through OS X, but as discussed OS X was now powering iPhone. The framing memo was not a plan nor was it even a plan for a plan, rather it serves to bound the release and help encourage the team to focus on specific technologies, competitors, and challenges. There was much for the team to collectively learn. At the time, I summarized Apple competition as follows:

Apple.  Apple OS X is a very strong product coupled with amazing hardware.  The transition Apple has made from OS 9 to a modern OS architecture on Intel hardware is on par with the transition we made to both the NT code base and to 64-bits.  From an OS technology perspective we see the strength of OS X among universities and administrators who find the Mac (Mach-based) kernel and associated shell, tools and techniques “comforting”.  From a user experience perspective, we cannot be defensive about the reality that Mac hardware + OS X + iLife continues to be the standard by which a PC + Windows 7 + Windows Live will be judged in terms of technology, and then [sic, how] the purchase experience, post-sales experience, and ecosystem have grown to be considerable strengths.  While we have only some details, the hygiene work being done on Snow Leopard is likely to generate significant positive views as the OS becomes “leaner and more streamlined” and likely claims about being more modern with respect to graphics, 64-bits, and user-interface.  As we describe below, the sharing of code and architecture, particularly for important strategic elements of the OS, between the iPhone OS and Leopard is technically interesting and certainly responsible for some elements of the platform success.  Apple is not without blemishes, but in planning Windows 8 we must focus on their strengths and assume they will continue to execute well on those.  Being deliberate and informed about competing with Snow Leopard and, relative to iPhone, making sure we build on the assets of Windows for our Windows Phones should be strongly considered.

Rather than describe Android directly, I chose to consider Android as another variant of Linux. At the time, the only Android phone was the convertible Blackberry-like device, the HTC Dream. Android 2.0 with full support for multitouch would not release until Windows 7 launched to market about three months later. Note the caution about Google control of Android expressed even then—it was obvious this would be a tension point down the road. In this context the competition with Linux also included desktop Linux:

Linux.  Like so many competitors we have faced as a technology company, just as we thought “this one won’t reach critical mass” we saw two events provide a breath of life into “Linux on the desktop”.  First, the rise of Ubuntu, both the technology/packaging and the “movement”, has created a rallying cry for OEMs and for the Linux community.  Second, the low-cost PCs that initially came with Linux (due to the footprint “requirements”) created a new incentive for OEMs/ODMs to find a way to make it work.  While we were effective in establishing a value proposition for Windows XP on these PCs, the seed that was planted will continue to be revisited by PC makers and designers.  They appreciate the potential for business model innovation, componentization, tailoring, and also the opportunities to differentiate using the open source aspects of the OS.  It is worth noting the reality that Linux on the server continues to dominate in many important workloads and the server plans are going to inform the planning of the base OS work such that we are extremely focused on defining specific scenarios where we make significant competitive progress with Windows 8 against Linux on the server.  Within the context of Linux it is especially important to call out Google Android which will likely be funded by Google for some time and represents an OS choice for mobile phones and phone architectures encroaching on the PC “from below.”  Android can provide OEMs with the opportunities similar to Ubuntu, however Google is walking a careful line in providing Android where they can possibly lower costs for OEMs, or even subsidize the bill of materials for a device, counterbalanced by OEM wariness that Google will take too much control of additional revenue streams.

Perhaps most importantly were the two “non-traditional” competitors: “Browsers” and “Phones/Alternative Architectures.” The rise of Google Chrome was proving highly problematic, not just because of the loss of share, but because of Google’s determination to add anything and everything from a PC to a cross-platform browser runtime, which is still going on to this day.

The view of the phone is much more interesting considering how the world evolved. Speculating about Android would prove accurate in just a few short months. I was deeply concerned about the PC ecosystem and the potentially rapid convergence of PC and phone hardware in the direction of phones, as well as Apple’s unified OS strategy. This was in stark contrast to where Microsoft began its mobile journey a dozen silicon-years earlier and where it was today.

Phones and alternate architectures.  The iPhone is referred to by many as a new form of portable computer—“it has a real OS” many have said in reviews.  The Google G1 phone running Android is likely to be made available in more PC or “handheld” form factors beyond the single-handed in-your-pocket screen.  The desire for the ultimate device that has the power and capabilities of a PC along with the convenience of always on wireless connectivity is beyond alluring.  It is what we all want as consumers.  To deliver on such a vision many might say that Windows can never power such devices—perhaps that is a statement about “business model” or a statement about technology.  Sometimes it is just competitors declaring “[Big] Windows will never be on a phone”.  The reality is there was a time in the history of Windows phones (CE, PocketPC, etc.) where the synergy between “little” Windows and “big” Windows was technically robust in reasonable ways.  For a variety of reasons we diverged.  Today we have hardware designers and phone company customers facing a choice between the OS that supports phone networks, voice, and other “phone” scenarios super well, but does not have the rich ecosystem support of Windows 7 and runs on a small set of hardware chassis, and Windows 7 running on a mainstream hardware platform with broad ecosystem support and openness.  These platforms differ in bill of material costs, power consumption, and so on.  Much like the “browser is all I need” statement there is a significant amount of extrapolation that both excludes Windows and assumes the Windows platform cannot compete.  Our chip partners for Windows are working hard on bringing the x86 architecture “down” and we need to be there with Windows software.  And at the same time we will strongly consider how to run Windows on alternate hardware platforms and learn what that would entail—we will work to bring “big” Windows to mobile chipsets, but we have significant groundwork to do before we know the practicality of such an investment.  Our Windows 8 planning will most certainly take into account the role of sharing new code across Windows and Windows Phones, starting with the latest Internet Explorer as we are already working on.

There were several technology bets in the framing memo, putting stakes in the ground for how we would think about significant efforts on the deeper technical challenges of the release including a big effort to evolve the PC. While I don’t want to fast-forward, these abstract definitions are what will lead to the work on ARM processors (see emphasis above), a new application model, and ultimately first-party hardware.

The technology bets on evolving the PC included:

* Mobile Devices

* Converged OS between Windows PCs and phones

* Mainstream PCs and the market turn to laptops

* High-end PC form factors such as meeting room PCs

* Shared computing PCs such as we were seeing with remote desktop and cloud computing

Taken together with the work on cloud, the memo defined the “Modern Windows Experience” as including the following (the original even included a fancy Office Art depiction). The use of “modern” will become a touchstone of the release and how we describe Windows 8:

* Base operating system: the OS kernel, device management, connectivity, and storage

* Graphics, Presentation, and Interaction: the visual aspects of Windows including the APIs used by applications and the end-user experience

* Browsing: the technologies required for browsing the web, but also how those will be reused across the platform to build rich applications

* Windows “Form Factor” Experience: tuning the experience for different types and sizes of computers from handheld through laptops and desktops to servers and alternate form factors such as tablets and wall computers

* Windows Live: the broad set of cloud-based services that complete the Windows experience including identity, communications and messaging, storage, and sharing

As with past releases the framing memo went out to the whole Windows team. The bulk of the team was super focused on shipping Windows 7 and mostly appreciated the informational aspect of the memo. The real work for program management began with JulieLar’s “Windows 8 Planning Memo” which she sent to the entire team just after Windows 7 RTM for most worldwide markets. The purpose of the framing memo was to establish the “bounding box” for Windows 8 and bring together the best of top-down, bottom-up, and middle-out planning. The framing memo was the top-down step. The planning memo develops potential themes for the release and explores those.

Julie outlined the following planning themes which would then be the structure used to facilitate brainstorming, prototyping, and then ultimately the product vision:

Blending the Best of the Web and the Rich Client. Recognizing the declining role of Win32 development and the increasing dominance of web development and associated tooling, this area asked to develop solutions to building an engaging and useful platform for developers. What kinds of apps do they want and how would they be distributed? How could the capabilities of Windows enhance what we see as limited web applications?

Defining a Modern PC Experience. Julie wrote, “The basic elements of today’s Windows user experience—the Desktop, Taskbar, Start menu, and Explorer—were introduced in Windows 95, and their success has made Windows the world’s most familiar computing environment. But today’s modern world is in many ways different from the mid-1990s world in which Windows 95 was designed.” This area asked nothing less than to bring the Windows user experience up to a level that would support existing scenarios and provide a better solution for the next billion customers who see the smartphone as their first and potentially primary computing experience.

Extending the Reach Of Windows. The Windows model of licensing and OEMs served us extraordinarily well, but with the rise of smartphones there were new revenue opportunities and sales models for Windows that can be supportive of reaching the next billion customers. Key among these was the role of mobile operators (“telcos”) in the sales process.

Connecting to Windows From Anywhere. Typically, the PC of 2010 operated in a world where everything is on the PC except what is viewed through a web-browser. PCs could be greatly improved by making use of cloud services available with Windows Live to make all your files, settings, communication, and collaboration easy from any PC or device. What role could the cloud play? Synchronization across devices was something the company has tried many times before though we had limited success. How could it work this time?

Helping IT to Deliver Work Anywhere Infrastructure. Most of Microsoft’s revenue and much of the upside opportunity came from doing a fantastic job for enterprise IT. This theme asked questions about deployment, security, encryption, and even the mundane such as how to easily replace a lost, stolen, or broken PC.

Showcasing Quality the First 30 Days and Over Time. This area asked what have historically been the most intractable of questions about Windows. “The burden of ensuring that a new PC is running as well as it should is placed on the customer who purchased it.  As a result, the first days of usage, rather than being a period of exploration and fun, were often prove to be labor intensive and exasperating.” Julie added, “Making matters worse, Windows itself is not running at its best during the first days of ownership.  …Windows performs all of its initial self-tuning and post-out-of-box-experience tasks during this critical time.”  When compared to what customers were starting to see with smartphones, Windows looked downright archaic.

Synthesizing these together into the vision was the next step. While most of the planning team were brainstorming and prototyping, we had some significant engineering work going, the success of which would prove crucial in enabling our ability to achieve a breakthrough Windows 8, such as working out the feasibility of Windows on ARM.

A bigger worry, did we set out to do too much? We knew that the team would naturally gravitate to ideas that made the PC better in incremental ways. We also knew some people would push hard on extremely bold ideas. The risk for any large organization is obvious…trying to do too much. The less obvious risk…ending up with a plan that is a too much of the old and a little bit, but not enough, of the new. Inertia is one of the most powerful forces in a large company.

I was worried. We only wanted to spend a few months “just” planning, still an enormous undertaking. Windows 7 shipped in August, with this planning memo and associated work taking most of the Fall. It was looking like we would start the project in the Spring of 2010, which meant an aggressive RTM of Windows 8 in the summer of 2012. Still, that would be exactly three years between releases, something never accomplished before. Promise and deliver.

On to 098. A Sea of Worry at the Consumer Electronics Show

In the era of “boxed” software release to manufacturing was a super special moment. The software is done, and the bits permanently pressed onto a DVD disc. That disc, the golden master, is then shipped off physically to duplicators around the world and then combined with another artifact of the era, a box or in the case of Windows 7 a plastic anti-theft DVD contraption. While Windows 95, the excitement of computing and the newness of internet set a high-water mark for launch events, the completion and launch of Windows 7 was a major worldwide business event. The industry was looking for optimism as we emerged from the Global Financial Crisis and the ensuing slump in PC sales. Windows 7 was just the ticket and the launch would prove to be part of a massive uptick in PC sales or as some hoped a return to ongoing up and to the right curves. But could that really be the case?

Back to 095. Welcome to Windows 7, Everyone

The months after the PDC were extremely intense. We had set out to promise and deliver, but the success of the PDC had managed to inflate expectations. These were not false expectations—the use of the product was widespread and broadly satisfactory. That success is what raised expectations. PC makers, Wall Street, OEMs, and enterprise customers knew the product to deliver and were not just impatient.

We made a significant number of changes from M3 to beta. With our improved engineering system changes were made in a controlled though collaborative manner. Each change was discussed by many people and then the code change reviewed—no holes punched in the wall. With each passing day it was more difficult to make changes while we aimed for stability of the beta. The most important thing about shipping a beta is not that it is perfect but that it ships in a known state. If something isn’t right that’s okay, as long as it’s known. In the case of Windows 7, we knew work and bugs remained but were highly confident that millions of people would try out the beta and have a great experience.

That methodical crawl to beta went on for weeks, each day making fewer changes and calmly making it to sign-off. Then it was time to ship the beta.

On January 8, 2009, at the Consumer Electronics Show in Las Vegas, SteveB announced the availability of Windows 7 Beta. The venue and the announcement from the CEO made this a significant worldwide event. It was covered on CNN, BBC, and more. That was exciting and even felt a bit like old times for a moment.

I watched from back in the green room because we were getting ready to turn on the web site for download and had no idea what to expect. While the internet was old news, downloading a gigabyte DVD image was hardly routine, especially from home and not something the internet was yet equipped to handle reliably.

To have some sense of control, we set a limit of 2.5 million downloads. Back then, before everyone had gigabit internet at home, a massive gigabyte download was something that would stress out the internet. As the keynote was going on we watched downloads begin. They quickly reached our limit while the keynote continued. A few calls to Redmond and we removed the throttle and began to rewrite our press releases with ever-increasing numbers. We extended the beta downloads through the end of January and had many more millions of installs than downloads as the download made it to all sorts of alternate and backup sites. We also learned a lesson in distributed computing that day.

For the beta we issued unique product registration keys which became the scarce resource. We soon removed the limits on activating those keys as well. While the download site was structured to choose 32- or 64-bit along with locale to then generate a key, many figured out the URL that went directly to the 2.5GB download and passed that along. We just didn’t want to be overwhelmed with Watson and SQM data so capped the release at 2.5 million. That was silly but at least we received an indication of the excitement. There was a lot!

Every day we tracked bugs with Watson and observed usage with SQM. Hardware vendors were providing updated device drivers that were anxiously downloaded by millions of testers, many seeing new drivers arrive automatically by Windows Update. More new PCs would arrive to be qualified. More legacy hardware would be retested. More of the over 100,000 apps in the wild would be checked for compatibility. More enterprise customers would tell us that they were anxious to deploy Windows 7.

Many reviewers chose to review the Beta as though it was final or at least something regular people might care about. It would be easy to gloss over this but for me it was an important part of promise and deliver. It had been a very long time, perhaps never, when a first beta for Windows was considered broadly usable and also had customers asking if it was okay to deploy even more broadly. Promise and deliver.

David Pogue, hardly a fan of Windows, practically filled an entire page of The New York Times with his review “Hate Vista? You May Like Microsoft’s Fix” where he concluded “For decades, Microsoft's primary strategy has been to put out something mediocre, and then refine, refine, refine, no matter how long and no matter what it costs, until it succeeds. That's what's exciting about the prospect of Windows 7. It's Windows Vista - with a whole heck of a lot of refinement.” Microsoft was back to making sure it got products right.

In The Wall Street Journal, Walt Mossberg said in another front page of the business section review “Even in beta form, with some features incomplete or imperfect, Windows 7 is, in my view, much better than Vista, whose sluggishness, annoying nag screens, and incompatibilities have caused many users to shun it. It's also a serious competitor, in features and ease of use, for Apple's current Leopard operating system.” He also posted a video review as part of some of his more recent work creating video reviews.

All we needed to do was finish.

By July 13, 2009, build 7600 was pronounced final and signed off on by GrantG and DMuir, the test leaders for WEX and COSD.

Windows 7 was ready for manufacturing.

July 9, 2009 was also my 20th anniversary at Microsoft and the team help me to celebrate by dressing like me—jeans, v-neck sweater, t-shirt. At least I was predictable.

Shortly thereafter at a sales kickoff in Atlanta, the annual MGX, we surprised the global field sales force and created a media moment when SteveB, KevinT (COO), and I held up a “golden master” DVD (a gold DVD), symbolically signing it as we announced that Redmond had signed off on the Windows 7 RTM build. It was a release that 10,000 people worldwide had contributed to and would likely end up on over 1 billion PCs. It was complete with another photo of me looking uncomfortable celebrating with SteveB on stage at the sales meeting. The sales meeting was always a country-mouse/city-mouse moment for me.

Just as soon as that excitement was over, Microsoft announced what turned out to be the worst earnings in corporate history with a 17% drop in quarterly revenue in July 2009 the end of the fiscal year. The tough part was we knew this was coming while we were on stage, which made our celebration much less about the past and more about hope for the future. While many would blame the economy or Vista and some would even cite the recently announced Google Chrome OS (the predecessor to the Chromebook), the truth was much more secular in nature. PC makers were struggling on the bottom line as the 40 million netbooks as exciting as they were lacked profit. The astronomical rise in netbook unit sales discussed in the previous chapter led many to assume a bullish future. In fact, netbooks masked a secular decline in PC sales. We would know more as the year progressed as new PCs were sold with Windows 7 during holiday season 2009.

After the celebration, the team collectively exhaled. It was August and time for vacations, but we didn’t have a lot of time to waste. Come September, we had to get the team fully engaged to plan what was coming next or it would be a massive effort to regain momentum.

Our team of administrative assistants outdid themselves with a wonderful ship party held on the activity fields in Redmond. In contrast to other Windows events, I would say this one was less eventful and even comparatively subdued, but still enormously fun for the team. We had custom cakes and cupcakes, tons of food, family-friendly games, craft beers contained to a two-drink limit within an enclosed area, and even Seattle hometown favorite The Presidents (of the United States) as the special musical guest. They took their popular song “Cleveland Rocks” and wrote the lyrics as “Microsoft Rocks.” I still have a bootleg recording of that.

The competitive issues we were facing weren’t going away. The organization was about to change and clarify responsibility for dealing with those.

Launching Windows 7

We had a fantastic foundation to build upon and all we needed to do was deliver an odd-even result—meaning a good release after the Vista release—and each of our core constituencies would breathe a sigh of relief. Looking outward, however, it was obvious the world was a very different place than when we started Windows 7. It was not clear any of those constituencies or even our own team were prepared.

There was still a product to officially launch, but not before some realignment at the top of the organization.

Bill Veghte (BillV), who had started at Microsoft right out of college and had overseen Windows 7 marketing through to launch planning, decided after two decades with the company that he wanted an opportunity to run a business end-to-end and announced his intention to depart Microsoft. After a transition he would join Meg Whitman at Hewlett Packard. With that, SteveB wanted to put all of Windows under one leader and asked me to do that. He really wanted to elevate the job title which I pushed back on because of the way Microsoft was structured (and remains so) we did not really have what most consider true ownership of a business. Nevertheless, that was the origin of the job title of divisional president.

One of my first tasks was to hire a marketing leader to take over from BillV, one who would best represent the collaborative culture we aimed to create. I wanted to bring finance and marketing together under one leader because the Windows business uses billions of dollars in pricing actions to fund marketing through OEMs. Tami Reller (TReller) had been the finance leader for the Windows business, reporting to the corporate CFO. When she joined Microsoft 10 years earlier following the Great Plains acquisition where she had been leading marketing. I got to know her then as the acquisition fell under my then manager, Jeff Raikes (JeffR). She was the perfect combination of marketing and finance leadership for a business where those went hand in hand and brought a great deal to our management team.

Microsoft wanted (needed) a big launch for Windows 7 and so did the industry. As had become a tradition for me, I wanted to spend the launch in Asia while my peers led the US event. My connection to my Microsoft family in Japan, China, and Korea ran deep and the business for Microsoft in those countries was huge. I couldn’t be in all places at once, so I chose to attend the launch in Japan. No one loves a retail launch more than Japan.

I arrived two days before the October 22, 2009 launch. I never worked harder at a launch event. From first thing in the morning (easy because of the time change) until well past midnight (well supported now with Modafinil) because of excitement at retail stores, we did press visits, interviews, broadcasts, met with customers, and more. We’d shuttle from event to event in a Japanese microvan—all of us in our blue suits and ties with a stack of name cards.

The above are two YouTube videos from Japanese Windows fans who recorded Akihabara Electric City the night of the launch as well as the Ultraseven appearance.

For years, even way back when I worked for BillG, I had been going to Yodobashi Camera in Akihabara (and Shinjuku) to see what Japanese consumers were buying and to buy assorted USB and power cables that are exactly the length I need. The size of the Yodobashi flagship is unimaginable. The evening before the event we got an incredibly cool behind-the-scenes tour of the store, getting a look at the entire operation at night as they prepared the signage that would blanket the store for our event there the next day, such as the big decals that covered the walls of the 5-story escalator. As someone who grew up in the shadow of Disney World, the underground tour of Yodobashi was much like the underground of Disney, and about the same number of people visit each year I was told.

The team put on an outdoor event at the front of the store the evening of the launch with all sorts of famous-in-Japan anime/cosplay actors and tech celebrities. And when the first copy was sold that evening, we did a press event right there at the front of the store. All along the main street of the Akihabara Electric Town, Chuo Dori, there were events in front of the many stores selling PCs and software.

Microsoft Japan, MSKK, had come up with a crazy promotional partnership with Burger King. The chain created a seven-layer Whopper to celebrate Windows 7. It was five inches tall (13cm) with seven patties totaling 1¾ pounds (791g) of beef. It was unfathomable, even for a no-carb, protein person like me. The first 30 customers got to buy the burger for ¥777, or about $9 at the time. The launch team and I snuck over to the Burger King around the corner from Yodobashi and ordered the monster burger. None of us could eat it elegantly or even try to finish it, but we got some hilarious team photos of the attempt and general celebration.

At a hotel ballroom, we held the main launch event for the press, featuring all the new PCs from Japan PC makers. The event featured an Ultraman theme. Why? I knew about the movies but was not a huge follower. What I learned was that Ultraseven was the third installment of Ultraman from the late 1960s. It was still wildly popular in some circles in Japan. The launch had a cast of people doing choreographed battle scenes in Ultraseven and Ultraman outfits. It was something to see. I filed this away for another Lost In Translation memory.

MSKK hosted a both a casual user group meeting and a formal business launch as well. At the user group meeting we did demos and gave away a bags of Windows 7 logo gear among a series of demo stations at a cool Akihabara exhibition space down the street from Burger King. I wore a super cool Windows 7 windbreaker which I still have.

The business launch was a formal ceremony highlighting the broad support of both hardware and software for the launch. Joining Microsoft was the head of Dell Computer Japan. Together with a group of MSKK employees and partners we participated in the traditional celebration kagami-biraki or cracking open a rice barrel with big wooden mallets.

I’ve had the privilege of experiencing many product launches in Asia, but this time, for Windows 7, it was next level. MSKK is a gem of Microsoft. When I am lucky enough to be in Tokyo, even years and years later, walking around Akihabara I have the warmest and most vivid memories of the launch and my friends from MSKK. And sometimes my stomach hurts a bit thinking about the Burger King, which recently closed just before the pandemic. The news coverage of the event in Tokyo which was amplified across the important Asian markets was wonderful.

Our confidence was high heading into reviews, which broke with availability of the product and new PCs in retail stores—we had plenty of positive reviewer experiences and no deep concerns. That’s what came from being not just in beta but running as the primary OS on reviewers’ and enthusiasts’ PCs for months. We risked a reviewer becoming somewhat bored or even cynical with the release simply because there was little new from the beta and no product drama to speak of. Hundreds of positive stories broke across print and TV. Local reporters did a lot of product reviews and buyers’ guides at that time. Waggener Edstrom worked tirelessly in the United States to feed them information and support.

Walt Mossberg’s review evoked a positive tone that started in January with the beta release. For his RTM review he said, “Bottom line: Windows 7 is a very good, versatile operating system that should help Microsoft bury the memory of Vista and make PC users happy.” The headline read, “A Windows to Help You Forget: Microsoft's New Operating System Is Good Enough to Erase Bad Memory of Vista.” There was little more we could ask for in a review.

Ed Baig of USA Today and one of the most widely read reviewers made it clear how positive he was on the product when he said “What you'll notice is that Windows 7 is snappier than its predecessor, more polished, and simpler to navigate. Screens are less cluttered. It has better search. Windows 7 rarely nags.…It sure seems more reliable so far.”

Windows 7 was the first major release of Windows not to double the requirements for memory and disk space. While the box maintained the same requirements (also a first) in practice the reduction in memory usage and focus on Task Manager paid off handsomely. As much as we were proud of the business success, the engineering success of Windows 7 was among the most significant in company history and the reviews reflected this improvement in core software engineering competency. JonDe brought his engineering excellence to all of Windows.

Major PC makers used the time from sign-off on the build to the October launch event to prepare the first Windows 7–ready PCs and get them into stores for holiday sales, including Black Friday in the United States.

Industry analyst firm Gartner declared the “recovery of the PC market on a global level,” with preliminary numbers showing a 22.1 percent increase over the previous year. Their quarterly analysis was effusive relative to their own reports just months earlier that were doom and gloom. More than 85 million PCs were sold in the fourth quarter of 2009, up more than 10 million units from 2008. This even though we were in the midst of a global recession. One year earlier, the top line was that PC sales had crashed. By the end of the first quarter, Gartner would upward revise their forecasts for 2010 to almost 370 million units, growing nearly 20%. The primary reason was that mobile computing, including netbooks, was on fire. Gartner concluded “It was the strongest quarter-on-quarter growth rate the worldwide PC market has experienced in the last seven years.”

It would be incorrect to assume cause and effect relative to Windows 7. There existed pent-up demand for new PCs to replace aging ones. Windows Vista had caused many, both at home and at work, to hold off buying new PCs, and the recession further slowed those decisions. Windows 7 brought many people back into the market. The shift to mobility was helping PC units, but the low cost of netbooks hurt the profits of the major PC makers.

The competitive forces were real. Apple was doing very well in the US (and Japan) and finished the year selling 24% more units year over year. The strength in consumer sales was the headline supported by the so-called consumerization of IT, where consumers were buying their own preferred PCs to do work rather than rely on stodgy corporate PCs that were slower, heavier, and burdened with IT software.

It felt, at least to me, that I’d been holding my breath for more than three years.

I walked through Meiji Garden and Shrine early the morning of my flight home, a travel day tradition and one of my favorite places on earth. The world’s economy was still in shambles from the Global Financial Crisis. The PC sales everyone was excited by were obviously juiced by the start of an economic recovery and by netbooks. These were not going to last. When it came to netbooks, the major OEMs were anxious to exit the market and return to their view of normal. The problem was, as it became clear, netbooks were additive to a shrinking market. Consumers wanted portability. They were willing to try netbooks, but the product could not meet expectations.

When NPR reviewed Windows 7 in a very positive review, even the introduction espoused the end of the operating system, saying:

We are in the modern world now and, while Windows continues to be the default OS, everyone is talking about Mac OS X, Linux and the second coming of, wait, no, just the much-anticipated arrival of Google's Chrome OS.

The future is the Web, not the OS, and everyone knows it.

As the Narita Airport customs officer stamped my passport and I walked through the turnstile, I could finally exhale. I think my stomach still hurt from the attempt at the seven-layer Whopper. Everyone else was heading back from the New York launch events and, other than the coverage I read, I don’t remember if we even took the time to share stories.

I had the same feeling I had when Office 2007 finished. As happy and proud as I was, it felt like the end of an era. With the huge shift happening in PC sales, PC makers, the internet and cloud, mobile phones, there was no denying we were in another era. When I looked at Windows 7, I did not have a view of “look what more we could do” as much as “we’ve done all we can do.”

What I do remember more than anything was talking to members of the team in the hallways, at meetings, remote offices, or over email throughout the course of the release. No matter what was going on and how difficult things got, I will always cherish all the people who shared their feelings about doing some of the best work of their careers—thoughts I still hear even as I write this. It was incredibly rewarding to hear. That wasn’t about me, but about the system and the plans put in place by the team of leaders we assembled. The Windows team was a new team. It was so ready to take on new challenges.

With RTM everyone on the team received their ceremonial copy of Windows 7. It is nice to have something to put on a shelf reminding each of us of the project and what we accomplished. For many, the next stop is the Microsoft Store to get upgrade copies for friends and family—another Microsoft holiday tradition.

The growth in mobility and demand for quality played right into Apple’s strengths, though not at first glance. The market continued to pressure Apple on low prices and did not see the weakness we did when it came to netbooks. On the heels of the Windows 7 launch, Apple released several new Get a Mac commercials. Among them was the segment “Promises” which reiterated all the times at new releases of Windows when Microsoft claimed the new release would not have all the problems of the old release. This one wasn’t accurate, but that didn’t matter.

In fact, Steve Jobs and Apple had a product in mind even more disruptive to the PC than the iPhone or MacBook Air…while we were just starting the next Windows release.

On to 097. A Plan for a Changing World [Ch. XIV]

While it is incredibly fun to do a first demo of a big product as described in the previous section, there is something that tops that and even tops the actual release to manufacturing. That is providing the release, actual running code, to a product’s biggest fans. It was time to welcome everyone to Windows 7 and put the code that the team had been working on since the summer of 2007 out for the world (of techies) to experience.

Back to 094. First Public Windows 7 Demo

Seattle summers are notoriously difficult on product development. After a long spell of clouds and rains, the beauty and long days of Pacific Northwest summers arrive, neither are particularly conducive to coding.

Summer wasn’t why I ended up here, but it certainly had an impact on me. On my first visit in 1989 during a dismal February, I saw the outdoor Marymoor Velodrome down the street from Microsoft and thought, this is going to be great. On TV it didn’t look as difficult to ride as I eventually learned it was. I sort of rode it exactly one time and that was the day my bicycle arrived from Massachusetts.

But alas, product development demands don’t end even with 15 hours of daylight.

It was going to get busy for the Windows 7 team. Our planned schedule called for the third development milestone to be complete by the end of summer 2008.

We were making progress, but the schedule was slipping. The code was getting better every week, but the overall game of schedule chicken that often plagued a large project was an historical concern. This was our first time as a new team going through this part of a product cycle. While we had a good deal of positive progress building a team culture, Windows was notorious for groups betting against each other’s schedules and being less than forthright with their own.

When HeikkiK was running the Office 95 ship room, he declared that everyone should be working to finish first, not simply to finish second to last. We needed to get to the end of the milestone as a team working together without looking for one group to blame, since it is never one group. JonDe and I had this same concern.

Everyone spent the summer installing daily builds on every PC we could. At one point, I must have had eight different PCs between home and office and was installing on all of them nearly every day. Every night I was installing a new build at home while doing email and other routine tasks. Even though my home “service level agreement” called for no beta software, an exception was made for Windows 7.

I was working at two performance extremes. I went to Fry’s Electronics and built my own “gamer” PC from the best components. I spent big bucks on a newfangled solid-state desktop drive (not common at the time), a crazy graphics card, fast memory, and the most ridiculous Intel chip available. I installed Windows 7. I was blown away by the speed (as well as the noise and wind emanating from the mini-tower). Starting Word or Excel seemed instantaneous. Boot took low single-digit seconds. It reminded me of the first time I used a hard drive on my father’s Osborne computer and how much faster it was compared to floppy disks. I used this PC when I sat at my desk at work, which wasn’t often as I was always walking around the halls.

At the other end of the spectrum were Netbooks. To the degree I could, I had taken a fancy to the Lenovo Netbook, the IdeaPad S10, and carried it with me everywhere especially at my favorite breakfast place (Planet Java) or lunch place (Kidd Valley) where I did a lot of Windows 7 blogging. Every Netbook was close to identical on the inside, but the Lenovo had a good screen and a rugged exterior. I modified mine, replacing the spinning hard drive with a then non-economic solid-state drive to better emulate future laptops like the MacBook Air. It was my primary PC for writing blogs posts, email, spreadsheets, and browsing, and the like, which was most of what I was doing. When we finally got to the Professional Developers Conference, this was the PC I held up with a bright yellow “I’m a PC” sticker on it, stickers marketing created in a jiu-jitsu move embracing the blowback from the Apple TV commercials.

I was constantly on the lookout for memory consumption and the number of running processes, the signs of bloat in Vista—fewer processes and less memory were better. Each process was a critical part of Windows. The number of processes had soared with Vista, and each had overhead in complexity and performance (in contrast to Linux, Windows processes were much more substantial and important to track.) Windows 7 was making impressive strides in reducing memory usage and process complexity. It served to make me feel connected to the engineering of Windows 7 and reminded me of counting bytes and seconds back in the day. I snapped screen shots of the Windows Task Manager and would bug JonDe and AlesH every couple of days.

Each day booting into a new build and seeing the progress was a great day. Each day revealed a crisis or challenge, but as a team everything continued to move forward.

Even though I was mostly an observer, the effort to improve performance was some of the best work of the release. It set a tone for making progress, but also for the ability of teams to work together. The conventional wisdom was that Windows Vista was inevitable and unavoidable as capabilities were added and the product grew which could not have been prevented. Windows 7 disproved that theory.

By midsummer, we had to slip the schedule based on our progress through M2 and M3. Originally our goal was to finish M3 and have a full beta in time for the previously scheduled Professional Developers Conference, PDC, in Los Angeles. We weren’t where we needed to be, so we took about an eight-week slip. The build at the PDC would officially be pre-beta, terminology we just made up. This would be our last slip. JonDe and I were privately relieved at the degree of the slip, but frankly the team was excited to be clearly on track, relatively, for the first time in many years. Depending on your experience or the context, eight weeks can seem huge or literally nothing. It was nothing.

SteveB sent a memo to all of Microsoft outlining some of the work to date for the whole of the fiscal year. The company had made a lot of progress on many fronts. The topic that had occupied a great deal of discussion, and was a good portion of his memo, was Google and competing with it on the consumer front and the potential relationship with Yahoo. SteveB also described the emerging cloud strategy, and the fact that more would be shared on that topic at our upcoming PDC.

Fiscal year 2008 was quite a year for Microsoft. Revenue broke $60 billion and operating profit grew 21 percent to $22.5 billion. The numbers were incredible. Still, the concerns about the PC and catching up on consumer services dominated Wall Street’s view. This memo was one of the early communications in a strategic shift to the cloud platform and you can feel the push-pull between cloud and the traditional model in the technology descriptions. It’s important to say that it was still super early in the journey to the cloud for enterprise computing and the topic was not top of mind for customers, especially as the financial crisis began to take hold. In fact, the feeling that the cloud was architecturally inferior to private data centers was by far the most common customer belief. Their future enterprise computing model was a data center running servers using virtualization. In 2008, the idea that there would be something of a new cloud operating system was mostly a view held inside the halls of Google.

In the memo, SteveB announced that KevinJo was leaving to be CEO of Juniper Networks, and that JonDe and I, along with Bill Veghte (BillV) leading marketing, would report directly to Steve, a reporting structure that remained in place through the release without issue. This was a standard and expeditious way to handle a managerial change at this stage of a big product. Incidentally, Satya Nadella (SatyaN) had recently moved to manage Search and ads in March 2007 and would also report to SteveB in a similar move.

In the lead-up to the PDC we began blogging publicly about Windows 7. With the focus on tech enthusiasts, IT professionals, and the trade press, I created a blog called Engineering Windows 7, or e7. An extension of how we thought about blogging for Office 2007, the blogs were the primary first-party communication channel for the product. We authored long and detailed posts, thousands of words, about the implementation choices we were making and how we measured progress. We offered tons of data to describe real-world Windows use (often my favorites posts). I authored posts but also introduced posts that other team members wrote, each expressing the design point of view and rationale. Many generated a great deal of dialogue and discussion and became news stories themselves. There wasn’t really a Hacker News yet for Windows coverage, but the comments sections of many stories read just like Hacker News would have read. Tech enthusiasts loved to dispute the data provided then just as today.

While to some press the blogging came across as a carefully crafted corporate message, nothing was further from the truth. We were simply blogging. The posts did not go through any corporate machinery or apparatus. They were as authentic as they could be. And the tradition worked so well that after the PDC it became a significant part of the communication of Windows going forward.

There were two relevant industry announcements that at any other time would have caused a great deal of distraction. The PC world was entirely focused on the PC, to the exclusion of the world of mobile phones, and, to some degree, browsers were still distinct as a challenge to Windows because they still ran on Windows and had yet to incorporate much beyond rendering text and graphics. Yet both phones and browsers would have announcements that would radically alter the competitive landscape for Windows 7.

In June 2008 at Apple’s WWDC (the World Wide Developer Conference, Apple’s version of the PDC), Apple announced the much-anticipated and predicted iPhone SDK and App Store which was teased earlier in the year. Initially, it had 500 apps, small relative to PC apps, but that number would grow at an astronomical rate. More importantly, it solved many key problems that had plagued PCs. In the announcement, which was a short note from Steve Jobs posted to Apple’s news site, he said “We’re trying to do two diametrically opposed things at once—provide an advanced and open platform to developers while at the same time protect iPhone users from viruses, malware, privacy attacks, etc.” This controversial change riled tech enthusiasts but also ushered in a new definition of computer, one that was safer and more reliable than anything a PC (or Mac) could offer. The emphasis below is mine.

Third Party Applications on the iPhone

Let me just say it: We want native third party applications on the iPhone, and we plan to have an SDK in developers’ hands in February [2008]. We are excited about creating a vibrant third party developer community around the iPhone and enabling hundreds of new applications for our users. With our revolutionary multi-touch interface, powerful hardware and advanced software architecture, we believe we have created the best mobile platform ever for developers.

It will take until February to release an SDK because we’re trying to do two diametrically opposed things at once – provide an advanced and open platform to developers while at the same time protect iPhone users from viruses, malware, privacy attacks, etc. This is no easy task. Some claim that viruses and malware are not a problem on mobile phones – this is simply not true. There have been serious viruses on other mobile phones already, including some that silently spread from phone to phone over the cell network. As our phones become more powerful, these malicious programs will become more dangerous. And since the iPhone is the most advanced phone ever, it will be a highly visible target.

Some companies are already taking action. Nokia, for example, is not allowing any applications to be loaded onto some of their newest phones unless they have a digital signature that can be traced back to a known developer. While this makes such a phone less than “totally open,” we believe it is a step in the right direction. We are working on an advanced system which will offer developers broad access to natively program the iPhone’s amazing software platform while at the same time protecting users from malicious programs.

We think a few months of patience now will be rewarded by many years of great third party applications running on safe and reliable iPhones.

Steve

P.S.: The SDK will also allow developers to create applications for iPod touch. [Oct 17, 2007]

The App Store also provided distribution and awareness to developers, a way to make money, and a way for Apple to vet apps that might be harmful to consumers. At the time the focus was on the fact that “30 percent goes to Apple.” When we saw the store, though, we knew the change would be monumental. To find software for a PC, the best someone could have hoped for was a website such as download.com. There existed varying levels of trialware, freeware, spyware, and malware. Apple had solved the software distribution problem, made sure software was reasonably safe and high quality, and given ISVs a huge new avenue for creativity and money. All on the most exciting computer around, the iPhone. That was the bad news. The good news was that the world still viewed PCs and phones as totally different things.

The world other than Steve Jobs and Apple, as discussed in the previous section. History would later reveal through email discovery the internal conflict that surrounded opening the iPhone to developers so broadly. The success would also go far beyond even what Steve Jobs anticipated.

All that I could think was, Time. I need more time

That was not all I thought. I also greatly admired what Apple was accomplishing. Like many who joined the Apps division in the 1980s, Macintosh was a crucial part of my early Microsoft work and the years before that. The old Mac ToolBox APIs are forever imprinted in my memories. Other than the die-hard fans, few generally acknowledged the consistent refinement and foresight in Apple’s software design. Several of us on the team were “original” Mac third-party developers from the mid-1980s and had always admired not only the results but the patience of their process. The continuous iteration and complete execution of what they did was admirable. Apple’s business in Macs was definitely not something we worried about, but their product execution was worrisome. I looked over their R&D spend and compared it to Windows and Office. In 2008, Apple for the first time eclipsed $1 billion in R&D for the year, a big uptick from the prior year, perhaps an indication of iPhone and iCloud ramp. The full Windows 7 team was spending about the same. There’s a huge difference in R&D when it comes to having a full ecosystem but at the same time R&D for hardware is much more expensive. The main point is not only were they building breakthrough products, they were doing so remarkably efficiently compared to Microsoft.

The Mac might have been a better product, but Windows won, and the winning product becomes the better product in the market. It was not until the iPhone and SDK came to be that the true appreciation of all they had done with so relatively little was so broadly understood.

A more shocking announcement that hit closer to home came six weeks later. Google announced the Chrome browser with a blog post and a classically Googley online comic that accidentally dropped too soon. Chrome, ironically named due of the absence of user interface chrome, would prove to be monumentally disruptive to the browser world. Google had dramatically improved performance and security in browsing compared to IE 7 from Vista (or IE 6 that so many were running still) and current leader Firefox. They had committed to open source and brought an entirely new level of energy to the browser battle. In his blog post, Sundar Pichai (yes, then product manager for Chrome) wrote in a nod to antitrust that the “web gets better with more options and innovation. Google Chrome is another option, and we hope it contributes to making the web even better.” In some ways, we were straight back to 1996 again. This would be a huge problem for the newly reconstituted Internet Explorer team, both immediately and going forward. Within a short time, there was a massive share shift to Chrome. Much like Gmail, Google released a product, seemingly out of nowhere, into what was viewed as a stable space. And they took it over. It would be years before privacy, tracking, and all the “evil” stuff Google’s browser would come to be, but at the time, a new competitive landscape was defined for IE. If our job was difficult before, it suddenly was even more so.

The PDC took place on October 27th at the Los Angeles Convention Center. Azure, the new name for Red Dog, was announced on the day one keynote. That proved to be both prescient and somewhat ahead of the curve for most attendees. Windows 7 was the second-day keynote and carried the bulk of the news for the show. In some ways the fact that most of the attendees did not seem to find Azure immediately useful made our jobs easier. Most attendees were still debating the proper way to pronounce Azure. The developer relations leader found the debate the night before particularly irksome as he was a Persian who had his own ideas of how to pronounce a word he claimed as native in origin. I was completely entertained by this late-night sideshow. Nevertheless, the fact that the attendees were somewhat puzzled by Azure compared to what they saw as vastly more interesting sessions on .NET, Avalon, virtualization, or the Windows 7 desktop. The disconnect was a harbinger of the disruption challenges the entrenched Microsoft would face.

While the audience for the PDC was professional developers there to learn the latest in APIs, tools, and techniques from Microsoft, the front rows of the main hall were the all-too-familiar members of the press. Looking out from the stage, I could see all the stalwarts of Microsoft beat reporters and technology press who had been frustrated by the lack of information of Windows 7. We were doing a keynote for developers who spent a few thousand dollars to be at the show, but in reality we were putting on a show that needed to be understood by the mainstream media and conveyed through the expertise of the industry press.

Steve Jobs had upped the stakes with his spectacular keynotes, increasing pressure across the industry to put on a good show. The normal Microsoft keynotes, the kind pioneered by BillG, were long and detailed with complex architecture slides and many graphics. These were somewhat enhanced as we moved to enterprise computing with obligatory and infamous “partner videos” featuring senior IT professionals in front of architectural diagrams or racks of equipment extolling the virtues of Microsoft’s strategy. The audience expected this type of keynote and expected us to write code on stage. By those measures, they keynote might disappoint.

While we tried to streamline the keynote, marketing insisted on having at least one customer testimonial. This is a pretty cool demo of work by Autodesk showing off the use of touch in Windows 7. (Source: Microsoft)

Having said that, as we planned for my first keynote leading Windows, I knew that the biggest mistake I could have made would have been to try to emulate what I was not. Most importantly, I also had to find a way to apologize for Vista without throwing the product or team under the bus. I had to find a way to be excited about Windows 7, realizing we had holiday PCs with Vista still to sell. Above all, our announcement was for a pre-beta, not even an official beta, though it was ultimately a distinction without a difference.

For my part, I went with who I was—like Sammy Davis Jr. used to say, “I’ve gotta be me.” The slides I showed were sparse and my words carefully chosen. While not one for grand entrances, I did choose “I Can See Clearly Now” by Johnny Nash as walk-on music. I was the only thing standing between the thousands of press and attendees, and seeing Windows 7 code. We knew people were there to see a demo, not a build-up or a long story. Just get to the clicks. I used just two minutes and not even 400 words then introduced JulieLar to step on stage and start clicking. I stepped down from the podium and remained upstage opposite Julie.

As soon as she brought up the screen on the monitor, people started taking pictures, some with their new iPhones, but most with Windows Mobile, and since they were live-blogging the event, we knew they were noting the build number that was visible at the bottom of the screen, confirming that our debate about what version of Windows we were working on would be part of the conversation. Within about a minute Julie got her first round of spontaneous applause and hoots. The demo was fantastic and every time she said “…works the way you want to,” we could feel the excitement. She demonstrated all features with both a mouse and by using touch on a monitor, including showing an on-screen keyboard with predictive text and more. The bulk of the demonstration emphasized “putting you in control” of Windows.

Once she was finished, I stepped onto the center of the stage and got to say, on behalf of the entire team, something two years in the making.

“Welcome to Windows 7, everyone.”

It was the perfect demo to introduce the product.

At some point in the keynote I needed address what everyone was waiting for which was what did Microsoft really think about Vista. While the press would no doubt take note, the credibility of what was said would rely on winning over the tech enthusiasts. More than any audience, the tech enthusiasts in the room were most disappointed in Vista and felt let down by the product. From March 2006 when I came to Windows, I promised to never be critical of what preceded me and I intended for that to be the case. It would have been so easy and so cathartic for the room to profusely apologize for Vista. It would have been equally wrong to pretend that we had not made some sort of mistake. I chose a path of subtlety and to acknowledge “feedback” in all its forms, including a few television commercials.  With a slide titled “Transition from Windows Vista” I framed the work we had done since Vista released as providing context for the day’s keynote:

As we set out to build this release of Windows, we really did have to recognize the context with which we were releasing Windows 7 and developing it. And that’s in transitioning from Windows Vista. We certainly got a lot of feedback about Windows Vista at RTM. (Laughter.) We got feedback from reviews, from the press, a few bloggers here and there. Oh, and some commercials. (Laughter.)

As part of the session, we wanted to highlight some of the features that were specifically relevant to the developer and enthusiast crowd. I took a moment to show seven features (the number 7 was used a lot) that were chosen specifically to generate applause for the crowd including: 1. BitLocker disk encryption (previously in Vista Ultimate), 2. Mounting VHD (a virtualization feature), 3. High DPI (support for really big monitors and normal sized text), 4. Magnify (an assistive technology for low vision that is also useful for product demonstrations), 5. Remote Desktop across dual monitors (the first live dual monitor demo we ever did), 6. Taskbar Customization (anything with customization is a pleaser), 7. Action Center Customization UAC/Notifications (the improvements over Vista for enthusiasts).

There was a clear call-to-action for developers including moving to 64-bits, using Windows touch, and more, but mostly to download and install Windows 7 pre-beta. People were doing that as soon as the lights faded.

We wanted the keynote to be easy and approachable, not usually the norm at the PDC. That also meant we would leave out a good deal of the team’s work and new features present in the beta build. To that end we created a massive “Product Guide” for the trade press. We would also follow that up with a workshop for them to attend where they would have a chance to ask questions directly of the product leaders. The full product guide ran 119 pages! The team promised and delivered, and you can see this from the prominence of the “Engineering Focus Areas” in the guide which were taken straight from the product vision and mock press release.

While we would not normally expect long-form reviews and deep dives in a pre-beta, Windows 7 was generating so much interest that the tech press was filing tons of stories as were individual bloggers who drilled into every aspect of change from Vista. YouTube was filled with demos created in short order. Windows 7 was the top of Techmeme, and not for messing up.

An example of the coverage was ActiveWin, a Windows-focused outlet, that wrote over 13,500 words plus screen shots on the pre-release. Andre Da Costa wrote the piece, releasing it on October 31 as the conference ended. They dove into seemingly every detail, even including their summary of the key goals of the release:

Key Goals:

* Under-promise and over deliver

* Reduce Compatibility problems and bring investments in Vista forward

* Reduce disk foot print and memory foot print

* Improve performance

* Secure, predictable

* Make the Windows and PC Experience easier

* Exceptional hardware and software support

* Bring future releases to market faster

* Personalized experience that defines you

* Superior mobility through reliable performance, power management

ActiveWin concluded better than anything we could have written ourselves. Promise and deliver.:

It’s safe to say I am overwhelmed, overjoyed and most of all excited about Windows 7. This is the release of Windows everybody has been waiting for, it’s what Vista was meant to be and beyond that. Windows 7 puts the user first; it’s about going back to the fundamentals of what an operating system must do. Managing and maintaining your PC is exceptionally seamless in Windows 7 and users will appreciate the tremendous improvements and advancements this update will offer on both existing and new hardware form factors in the future. Windows Vista set the foundation for a lot of what is happening in Windows 7 today. Windows 7 makes security Essential, but not aggressive like Windows Vista. The improved UAC will no doubt give consumers confidence in this feature, just the fact that you can tweak it to a certain degree is a welcome change. Businesses will appreciate the improvements to how the OS is managed and deployed while mobile users can get better experiences between their work and home environments. Home Networking has finally reached a level of ease of use that will make even the novice to make those PCs in the home talk to each other. There is still a lot of work to be done as this early glimpse shows. But Microsoft is on the right path with Windows 7, focusing on ease of use, compatibility, better ways of interacting with the PC and managing the personal data. This is an upgrade I am looking forward to and you should too.

Posts on Windows 7 experiences across all sorts of different hardware appeared. Engadget, everyone’s favorite tech blog, tried Windows 7 on an ASUS EeePC writing “just as Microsoft demonstrated, the relatively lightweight Microsoft OS required just 485MB of RAM when Windows 7 was fully loaded, sans applications of course. Hot.” The article’s title was even great for all the work the team put into this specific metric, “Lightweight Windows 7 pre-Beta on Eee PC 1000H looks very promising.” I can personally confirm that memory usage number.  

As a manager of a giant product and team there are, honestly, few truly rewarding moments that are also deeply personal. Nearly all the time there’s worry about how the team is doing and if they are finding the joy they deserve. October 28, 2008 was one of those exceedingly rare moments for me.

On to RTM…

On to 096. Ultraseven (Launching Windows 7)

In an era of huge software projects with a zillion new features in every release, there’s little more exciting than the first public demos. Such demos are also incredibly stressful to pull off. In addition to all the work to just get the code to demo-ready condition, there’s a lead-up to public disclosure, briefing reporters, and aligning partners. The first demo of Windows 7 was all those things and more, because we’d (or just I) had been so quiet for so long. This is the story of unveiling at least one small part of Windows 7 along with my own personal screw up along the way.

Back to 093: Netbook Mania

The second of three development milestones for Windows 7 was originally scheduled to end on March 26, 2008 which was eight months after the project start, Vision Day. We ended up finishing on May 9, which was a slip of 44 days. For any massive software project, this was fantastic. For Windows, it was doubly so.

It was even better than that. The new organization was starting to take hold. The product was emerging. The team was executing. We were building what we committed to build, and it was working. The “daily builds” were happening and by and large the team was running Windows 7 every day.

After two years in this role leading Windows, I finally felt like it would be OK to emerge and talk about what comes next. It is difficult to put into words the constant gnawing, sick-to-my-stomach feeling up until now wondering if we would deliver. We had definitely promised but for nearly 20 years I had seen leaders across the company say “the team is feeling good” or “we’re making good progress” or “the milestone is complete” only to see the project unravel or simply recognize it was never actually raveled.

For months I had been under immense pressure from OEM partners, our OEM account managers, enterprise account managers, investor relations, Intel, retailers, not to mention SteveB, and many more to just articulate a ship date or some plan. Hardly a week went by without receiving a forwarded email detailing the costs of not disclosing what we were up to.

Yet I was perhaps irrationally concerned that I would put something out there only to have to recant or adjust what was said. Many told me I was being overly cautious. Many said that it is better to open up communication and worry about having to correct it later. I just couldn’t shake the concerns. I felt Microsoft had one chance to make up for the issues with Vista.

Many perceived the Windows team was trying to become more like Apple and close off all discussion of a product until the moment it was announced. This was not the case at all. Windows is a different product, as described previously, and to bring it to market requires a huge ecosystem of support and that invests time and money. There’s no way to surprise the market with Windows because an entire industry needs to know about it, prepare, and execute to bring new PCs, peripherals, and applications to market.

For months, Roanne Sones (RSones) and Bernardo Caldas (BCaldas) on the Ecosystem team had been in deep technical discussions with partners about what would come next but had not yet committed to a timeframe. Any hints of a specific schedule (or business terms such as SKUs or pricing) would immediately make it back to the business side of the house and then to SteveB’s inbox. Even topics such as if there would be a 32-bit release (versus moving the ecosystem to 64-bit only) would have had broad implications for PC makers (and Intel). We had to walk a fine line between being excellent partners and creating an external news cycle that impacted partners as much as us. We knew that release dates were the most likely to be leaked, and the most damaging. Finishing a product with a giant, hovering countdown clock had dogged many past Windows releases. Yet, the partners needed time to prepare, and we were closer to finishing than starting. Windows 7 would soon be fully disclosed with the OEMs.

When asked in any forum, we said our goal was to release Windows 7 “within three years of Vista.” We were intentionally vague as to whether that meant release to manufacturing, available for enterprise download, first PCs in the United States, or some other market. Effectively, this gave us a buffer of about three months. And yes, that was sneaky, but it was the one concession I made to disclosure. I really hated that all people cared about was a date when a product was so much more than that. I understood, but still.

Then, in April 2008, BillG gave a speech, and inadvertently in one small part some believed he implied that Windows would finish in the following year. The press, who were there to hear about international finance at the Inter-American Development Bank meeting, ran with it and suggested Windows 7 would be ready much sooner than the previously planned three years from Vista. In fact, a year from April 2008 was sooner than our published schedule. That was not going to happen. Explaining that inaccuracy without stating the ship date was impossible.

It wasn’t just that Bill said the next Windows would arrive “sometime in the next year or so.” He also expressed his enthusiasm in what was certainly meant to be a throwaway line but came across to a tech industry desperate for any news when he said “I’m superenthused [sic] about what it [Windows 7] will do in lots of ways.”

We were close enough to completing the milestone that it was time to plan on officially talking to the press, who would be happy to talk off the record while also helping us to reduce the amount they would need to absorb all at once when it was time for stories to be written. In parallel the Ecosystem team began working with OEMs and ODMs on the detailed schedule and on software drops.

Our first stop, as it had been with every product I worked on since Office 95, was Walt Mossberg at The Wall Street Journal. Our meetings had become somewhat of a routine, perhaps for both of us, though by no means easy or predictable—I usually prepared an overly large amount of data to demonstrate how people were using our products out in the wild and hoped to both inform him while pushing for some positive recognition. Sometimes, yes, I went a bit overboard on the data. Walt was staunchly independent and would never say if I was persuasive, but he was always thoughtful in his questions and comments.

By this time, Katherine Boehret was joining Walt when he visited. She started with The Wall Street Journal out of college. By 2011 she had her own column called This Digital Solution, and also worked with Walt and Kara Swisher on the All Things D Conference (ATD). Katherine and Walt together were a formidable audience. They were both deep into products with their own unique perspective and would put up with absolutely no spin or marketing. They were advocates for their readers and strident in their desire to see PCs live up to their ease-of-use potential and played no favorites.

This meeting, about a month after BillG’s speech, had a dual purpose. We wanted to at least try to diffuse some of what they had no doubt perceived (rightfully) as a mess with Vista without throwing Vista under the bus, while also setting the stage for Windows 7. If all went well, we might even secure time at All Things D that year for a quick Windows 7 demo at the end of an already scheduled BillG and SteveB joint interview.

It was stressful. It was Walt. And Windows 7 was not fully formed for reviewers yet. Joining for the meeting or parts of it would be Julie Larson-Green (JulieLar) for Windows, Dean Hachamovitch (DHach) representing Internet Explorer, and Chris Jones (ChrisJo) discussing Live Services.

Meeting in a conference room in building 99 with a half dozen demo laptops on the table, I started with our usual printouts of data, showing them an overview of Windows Vista in market. Walt’s earlier review of Vista called it “maddeningly slow even on new, well-configured computers.” Katherine’s writings had been a bit less harsh, but not by much. I had to at least try to change their minds, but neither Walt nor Katherine was impressed. I took the time to talk about the landscape of PCs being sold and what was going on with laptops and Netbooks. In reviewing the original Asus Eee PC, Mossberg concluded it was a “valiant effort, but it still has too many compromises to pry most travelers away from their larger laptops.” That led to a hot topic for all reviewers, but especially Walt who had praised the MacBook Air: When Windows would see a MacBook Air competitor? Walt, JulieLar, and I had discussed the MacBook Air at the Apple launch event months earlier.

My lack of an answer on behalf of PC makers was not satisfactory for them, or me. As described previously, the PC makers were much more focused on inexpensive devices like Netbooks and not eager to take on Apple or the premium PC market.

Browsers were much discussed in the late 2000s, though not the one from Microsoft. We didn’t know it at the time but in hindsight it would be fair to assume they had been or were soon to be briefed on the forthcoming Google Chrome browser that shipped in late 2008. Still, Walt and Katherine wanted to know about Internet Explorer and privacy, a hot industry topic among a few, but especially them. We were woefully behind Firefox on core browsing capability, but we had a fantastic story to share about privacy features that DHach and team had developed, including blocking “tracking cookies.” We showed them how mainstream sites, like The New York Times, were doing a poor job communicating to users how much information was being shared and with whom, but with only vague permission or even disclosure. We did not go as far as offering ad-blocking which many tech enthusiasts would have appreciated, but we did plan on releasing and showed a “Do Not Track” feature.

During development, a series of meetings with lobbyists from the advertising industry discussing the Internet Explorer privacy features had led to veiled threats about anticompetitive behavior by Microsoft against ad-supported Google. Such hints or even threats were common from anyone connected to the Washington or government communities. This was unrelated to the Consent Decree, though there were still a couple of years left on that agreement and the oversight meetings that I routinely attended. As a result, Internet Explorer 8’s privacy features that were well received in this briefing would ultimately be scaled back due to an enormously frustrating push from the senior ranks of Microsoft’s legal department to capitulate to the lobbying groups to avoid drawing attention of regulators and to spare our own nascent advertising business from having to comply with privacy requirements. Do Not Track was essentially shelved even before we started. Today, the capability is a core part of Apple’s platform and the Microsoft Edge browser.

Our primary goal for the meeting was to showcase Windows 7. For the first time, we offered up a full disclosure of our overall goals and schedule. We trusted Walt and Katherine as we had built a great working relationship with them over the years, but, more importantly, because of their unmatched professional integrity.

After the requisite, but polite, reminder of the holes we had dug with Vista, we moved on to show some of the working features of Windows 7. After discussing Vista, Internet Explorer, and Live Services we moved to Windows 7 and the demonstration. JulieLar led a deep dive into our theme of “putting the end-user back in control.”

We discussed improvements to the dreaded UAC experience. User Account Control was introduced with Vista as a mechanism to lock down a typical consumer PC and prevent software from being installed by accident. Unfortunately, the swift reaction to such a “nanny feature” was universal loathing. It became a symbol for the dislike of Vista. As it would turn out, this feature was only the first of what would become the typical smartphone experience in years to come but being first at getting between tech enthusiasts and their downloaded software also incurred their wrath. It was also the subject of one of the more biting “Get a Mac” television commercials from Apple. Shortly after Vista launch, the internet was filled with instructions to disable UAC, which we definitely did not appreciate. Julie demonstrated the improved, though still secure, experience, which was much smoother and well-designed and added options for enterprise admins and tech enthusiasts to control the feature.

Julie’s demo succeeded in bringing together many concepts in the basic experience of launching programs and switching between running programs, and the array of distracting notifications and alerts. We were calling the collection of improvements to the Windows taskbar the new Superbar. With confidence, we compared the Superbar to the OS X dock, knowing we had solved problems that the dock had not.

We showed them the collaboration with PC OEMs on what would be new with Windows 7 PCs. The Ecosystem team had a long list of improvements to device drivers, supported hardware, and features to make the out of box experience for new PCs better for consumers.

And we had a surprise for them.

A big bet in Windows 7 was to implement a touch interface across the product, with features in the desktop experience and APIs for developers, as well as device and hardware management. We had been working closely with OEMs to define standards and configurations that would bring touch to Windows 7 PCs. OEMs were excited due to an entirely new engagement from MikeAng and team to enable quality touch in new PCs. They believed this would help differentiate from the Mac. We had an even bigger vision. We wanted this for all PCs eventually.

Months or more from broad pre-release and totally hush-hush, JulieLar demonstrated how we had moved applications from the original Surface table computer to PCs connected to desktop monitor touch panels. The Surface table PC, the original Surface, was a product developed in the Hardware division. It was not unlike an ’80s arcade table, featuring a modified version of Windows combined with custom hardware enabling a new form of multi-touch interaction. The table had found niche uses in Las Vegas, as information kiosks, and had been demonstrated by BillG at the previous year’s ATD Conference. As it related to Windows 7, there were touch APIs and the foundation of hardware support. Our main demonstration was mapping software that zoomed in and out using multitouch (like on the new iPhone) along with a virtual touch keyboard, which combined would offer up many opportunities for developers. On Windows, touch went beyond just using fingers but also included the digitizer needed for pen computing. It was the only feature BillG consistently pushed for in the release.

While touch was a part of Windows 7 from the start, there were two reasons we chose to emphasize it as an early Windows 7 feature in this meeting. Showing touch early was counter-intuitive because it was totally new and could have easily remained secret, for an actual surprise.

First, we wanted to garner broad OEM support for touch which was a long-lead feature for them. No OEMs were selling touch screens which meant sourcing and developing a product was a significant investment and effort. Momentum from the conference demonstration would represent a key public commitment by Microsoft.

Second, there had long been ongoing rumors that Apple would add touch to Macintosh and with the success of iPhone this seemed more likely. Whether such rumors turned out to be true or not, the opportunity to both garner ecosystem support and get ahead of Apple while also showing off a BillG pet feature while he appeared at the conference seemed positive all around. To BillG and other pen advocates, it seemed “obvious” Macintosh would gain touch and handwriting support. Microsoft’s Tablet PC was in market for years already and had not seen a competitive entry, so the logic went.

Neither Walt nor Katherine ever gave a thumbs-up reaction at a first showing, always reserving judgment until they used and wrote about a product themselves. Walt agreed to consider a demo of the touch features of Windows 7 at the ATD Conference a month later. They wanted to show more but we chose to keep the demo focused on what the ecosystem partners would value.

We had a lot of work to do, but we were nervous-excited.

With the ATD Conference pending, we were faced with a ticking clock, which meant we needed to disclose more details about Windows 7. The touch demo was too fragile and too elaborate to take on the road. We did not want to disclose details of the product without evidence, or, more importantly, a call to action for either developers or OEMs.

Adrianna Burrows (ABurrows after joining Microsoft) was the senior vice president assigned to the Windows account at the Waggener Edstrom communications agency. Adrianna drove the agency strategy for Office and was assigned to Windows when I moved. She was an astute communication and marketing pro, had a keen ability to create the right story at the right time, and was an elite distance runner and French speaker by upbringing. While she was at the agency, she was a key part of our senior leadership team. She was also the most competitive person I had ever known and would never accept second place.

People in communications rarely say not to talk when given an opportunity, at least that was the case in the 2000s. Reporters are going to write even without first-party commentary, and eventually whatever they write becomes more plausible than anything a company might later report. I had been quiet for too long. We were on the cusp of having a narrative created for us—one that would read something like: Windows 7 is going to be a “minor” service pack rushed out the door to fix the woes of Vista, built on a smaller kernel, MinWin, as the key technology. While that might introduce some compatibility concerns, it would enable finishing the release in early 2009.

Adrianna proposed a long form interview with a highly regarded Microsoft beat reporter, Ina Fried of the influential CNET. Ina was a thoughtful journalist with a wide-ranging understanding of the dynamics of the industry. She was widely read and by the right people. Adrianna was able to arrange to have a full transcript of an interview published along with Ina’s story to reduce the risk of being edited. I thought that was a solid idea at that moment.

Adrianna created the perfect opportunity for us even though I didn’t know what to say. More accurately, saying nothing was my comfort zone. While I never speak unprepared, I just did not work out answers that sounded credible for the questions I was obviously going to get asked.

I got on the phone with Ina, Adrianna right there with me in my office with the call on speaker. For an hour I did my best Muhammad Ali rope-a-dope. I acted as though I had been forced to make the call. I gave a lot of non-answers. I’m sure Ina was confused since we had initiated the interview. Adrianna was tensing up the whole time—I could see her eyes widen with each non-answer. The more I spoke, the deeper the hole I dug. My answers got shorter and my deflection increased—all I could think of was that I didn’t want to talk yet because I was so unsure of what we would get done and when. I could not figure out why I was talking and what the call to action was for readers.

I was trapped. I felt like we talked for the sake of talking and lost sight of the lead up to the first demo as the purpose.

Ina’s story ran the day after the call, right before Memorial Day, as we were heading out to the ATD in Carlsbad for our first public demonstrations of Windows 7. It was 3,000 words of me saying nothing.

The headline said it all: “In an exclusive interview, Steven Sinofsky offers up a few details on the new operating system and the rationale for why he is not saying more publicly.”

Adrianna wanted to punch me. I had blown an opportunity. I felt bad, but the damage was far worse for the team, who were confused because the interview ended up pushing the needle back to opaque from translucent. I made a mistake and handled it wrong.

I learned the hard way that I should have either not done the call or done it well.

Fortunately, All Things D gave us a chance to undo the damage. Bill Gates and Steve Ballmer were to appear on stage together for one last time. The goal for Microsoft was to show an orderly turnover as Bill announced the end of the two-year transition from Chief Software Architect to non-executive Board Chair and would no longer work day-to-day at Microsoft. After questioning, they would turn the stage over to a “surprise” demo of Windows 7 from JulieLar.

Julie and a veritable force of a dozen people had been at work hardening the Windows 7 demonstration for ATD. All had been setting up the demo since the night before.

On stage, BillG and SteveB discussed the transition answered and questions about what would happen in a post-BillG Microsoft. Steve describes the early financial controls and conservative hiring approach Bill put in place that became the hallmark of Microsoft. There is a touching and relaxed retelling of the way Bill recruited Steve to join the company, including Steve’s recollection of “a computer on every desk and in every home.”

Later, in a pointed question, Walt asked Steve, “Is Vista a failure? Was it a mistake?”

“Vista is not a failure and it’s not a mistake,” SteveB said. “Are there things that we will continue to modify and improve going forward, sure. With 20/20 hindsight, would we do some things differently?” He told Walt and Kara undoubtedly, yes, but then added that Vista had sold a lot of copies. (The video below starts at this clip.)

Walt asked if Vista had damaged the Windows brand. Bill jumped in with, “Well, there’s no product we’ve ever shipped, including Windows 95, that was 100 percent of what I wanted in the product…. We have a culture that’s very much about, ‘We need to do better.’ Vista gave us a lot of room for improvement.” The audience, and especially Walt, laughed.

Then Windows 7 was up.

JulieLar walked on stage and did a slick, six-plus-minute demo. It was the product that we had always envisioned, executed from an off-the-shelf laptop as well as from a desktop with a currently in-market touch monitor running Windows 7 software. It was live and that was terrifying for all of us. Notably, the code was barely working—clicking or tapping in the wrong place could have been a disaster. Still, it was a smooth demo.

Walt and Kara were constantly reaching over Julie’s shoulders and touching the screen to see what would happen. We had agreed to the scope of the demo and that we would not venture off and show or talk about other features.

Julie drew using a touch version of the venerable MS Paint and whisked through photo management, including “features anyone with an iPhone would be familiar with, such as two-finger zoom and slideshows.”

At one point, Walt noticed that the taskbar (the Superbar we showed off at our HQ meeting previously) looked a bit different and asked about it. Julie replied, “You know we’re not supposed to talk about that today.”

The mapping application from Surface Table was also shown but on Windows 7, including the live data for the Carlsbad, California, hotel we were in. The demo wrapped up with the playing of a multitouch piano application, which by coincidence was like one making the rounds on jailbroken iPhones. There was still no app store yet, but the technically savvy crowd figured out how to use the released developer tools to build apps and sneak them on to an iPhone.

Our demo was a success. Phew. Windows 7 was out there, at least in words, pictures, and videos.

The next step was getting pre-release code into the hands of developers.

On to 095. Welcome to Windows 7, Everyone

The Windows team was plugging away on Windows 7. The outside world was still mired in the Vista doldrums. Then in the summer of 2007 there was a wakeup call in the announcement and shipment of a new type of computer from upstart Asus, called a Netbook, a tiny laptop running Linux and a new chip from Intel. Would that combination prove to be a competitive threat or a huge opportunity for a PC world fresh off the launch of the iPhone?

Back to 092. Platform Disruption…While Building Windows 7

When a project like Longhorn drags on, the business is going to miss important trends. The biggest trend in computing in 2005-06 was expanding the PC to the rest of the world, something Microsoft and others called “the next billion” as the existing computing model reached approximately one billion of the world’s 6.5 billion people.

To outfit the next billion, many believed a new type of computer was needed. They were right. Many places where we would have liked to bring computers to the next billion lacked reliable electricity, air conditioning or heating, constant high-speed internet connectivity, and often had dusty environments as in Africa and much of Asia where I happened to have some of experience.

At the MIT Media Lab, Nicholas Negroponte, the lab’s founder, spearheaded a project called One Laptop Per Child, OLPC, that launched at the Davos forum in 2005. The rest of the world would come to know this as the “$100 laptop” at a time when most laptops cost about $1000 or more.

The price of $100 seemed absurd given that was less than an Intel processor and only marginally more than Windows, and that was before the rest of the hardware. Therefore, the initial designs of the OLPC would ship without commercial software from Microsoft or hardware from Intel. Instead, partners from anyone but Wintel lined up to help figure out how to build the OLPC. Almost immediately, Microsoft and Intel were blocked out of the next billion.

The resulting device, a product of some extremely fancy, perhaps fanciful, industrial design was a key part of drawing attention to what became known as the OLPC XO-1. The device featured several technical features that were aimed at solving computing needs for children in remote areas while also addressing the goal of ultra-low cost. The software was open source with a great deal of influence from historic MIT projects aimed at learning computing and programming. The effort even created a non-profit where people could go to a web site and donate the price of a device to have one distributed. The OLPC XO-1 was so cool looking that many people wanted one for their own use, right here in the US. The rollout and communications were exactly what you’d expect from the Media Lab—exciting and broadly picked up.

Microsoft through the Microsoft Research team and Craig Mundie (CraigMu) leading advanced technology spent a great deal of effort attempting to insert Windows into this effort. The company made progress but at the expense of causing some well-known members of the OLPC project to resign once it became clear that proprietary software was involved. Microsoft for its part would embark on the creation of a version of Windows that was ultra-low cost and stripped of many features as well, called Windows Starter Edition.

If there’s a theme in this work, both from Microsoft and MIT, it is that the core idea was that to bring computing to the next billion, products would need to cost less out of necessity and therefore they would need to do less and be less powerful. Often in the process of doing this the products would also go through transformation to make them easier to use, because apparently that is something required too. This is a fundamental mistake made time and again when addressing what the financial and economic world call emerging markets. Individuals in emerging markets do not want cheap, under-powered, or worse products. They certainly do not need products that are dumbed down. In technology, there is really no reason for that as products do get less expensive over time. Yet in the immediate term companies get stuck with their existing price structures and economics. And people in emerging markets are not less smart, they simply do not have access to the money for expensive products. I saw this dozens of times visiting the internet cafes in the most rural and economically disadvantaged parts of the world where students had no problems at all using a PC connected to the internet.

Microsoft would really get stuck in China where the limiting factor wasn’t hardware. People were buying huge numbers of PC components and simply assembling their own desktops that were on par with anything available from Dell or HP. They weren’t running Linux or Open Office, they were just pirating Windows and Office. The Windows team (I was still in Office) created a whole group to strip down Windows XP and add a shell to make it “easier” for emerging markets, again a dumbed down product. These changes to software were as much as way to make products favorable to the new markets as they were to make the product unfavorable to existing customers. Eventually, Microsoft came up with a plan to offer Windows plus Office to emerging markets, and China, governments for a very low price, so long as the computers were purchased by the government. At $3 per license this sounds like an incredible deal, but in all honesty was not that different from prices in many developed markets.

Still, the idea that the next billion required much lower priced computers, and somehow the rest of the world did not, would not go away. The need to serve this market drove the next wave of innovation from Microsoft and Intel, much more so than serving existing markets.

As Intel was mostly left out of the OLPC project, at the Intel Developer Forum in Fall 2007 they announced a new line of microprocessors with at least some emphasis on making lower cost computers for an expanded market. Intel demonstrated this processor in what is called a reference design, a PC made by Intel as a way to influence their customers to build similar PCs. The Classmate PC was a pretty cool looking laptop somewhat influenced by the Apple iBook, which brought the rounded edges, colors, and translucency of the iMac to a laptop form factor. Some would say the iBook itself owed its design lineage to the Apple eMate, based on the Newton, and sold to education markets in the 1990s. As a reference design, the PC shown could support a variety of screen sizes, storage capacities, and more, so long as it ran a new Intel low-power chip.

Also showing off the new chip in the 2007 Intel Developer Forum was the Asus Eee PC—that’s a lot of e’s, which stood for “Easy to learn, Easy to work, Easy to play” according to the box. The Eee PC was the first Netbook. It was also one of the smallest computers on the market. While there were many previous attempts at super small computers, such as the Sony PictureBook and Toshiba Libretto, there were years earlier and premium priced. This form factor and price were unique at the time. The first Eee PC 701 had the most minimal hardware specifications in any PC around and ran customized Linux and OpenOffice. It also had several games and entertainment apps.

The laptop was physically tiny as was the keyboard at 8.9 x 6.5 x 1.4 inches and under 2 pounds. It had a 7-inch display running at 800x480 pixels. For storage the Eee PC had only 4 GB of solid-state disk (like an iPod) and just 512MB of RAM. The price was $399 when it finally made it to the US, though the initial reports were it would cost $189-$299. It is worth noting that the same Intel chip was also available at retail stores in a laptop with a 14” screen, 80GB drive, a DVD drive, and 2GB RAM for about the same price. This reality would not distract from the cool factor or that it fit in any messenger-style bag.

The software load was kind of a mess, at least I thought so. The hardware, however, caught the attention of tech enthusiasts who were quick to turn the Eee PC into a tuner platform, looking to modify and replace components. Soon, modders were replacing the storage or adding more memory. There were web sites popping up devoted to modding the Eee PC.

The device sold quite well through the holiday of 2007 and that got our attention. So did the fact that modders were doing their own work to strip down Windows XP (from 2001) and squeeze it on to a 4GB system. One such modder was Asus itself which came to us wanting to officially modify Windows XP. There were three problems. First, they wanted Windows XP to cost the same as their Linux, which was $0. Second, they wanted to remove a bunch of XP just to make room on disk. What they wanted to remove was simply anything that took up space. It was kind of a free-for-all that reminded me of what enterprise customers did to Windows 3.x and Office 4.x back in the 1990s to squeeze on to 20MB hard drives.

The third problem was significant. Windows XP was done. We were over it. It was already seven years-old and we released Vista months ago. Vista was our bet. There was definitely no way Vista could be squeezed down, first of all. Second, Vista just went through the Vista Capable mess where the basic version of Vista became tainted in market and these chipsets were good enough only for Vista Basic.

If we began selling XP again for Asus, then we would have to offer it for every Netbook. And if we offered it for Netbooks then what would stop OEMs and ultimately customers demanding it for every PC. Suddenly it was looking like a roll back of Vista, especially if we participated. We didn’t really have a choice. Either we would lose the sockets to Linux, the modders would continue to pirate XP and hundreds of thousands of computers would be running a Frankenstein build of XP which already had tons of security problems, or we could suck it up and let the OEMs sell XP.

It turns out Intel was in a bit of the same situation. They were starting to worry that OEMs would want to make many more laptops out of the low power chips and that would take away sales of their more powerful and profitable laptop chips. Intel defined the Netbook category, and thus availability and pricing of low power chips, to require certain maximum screen sizes and configurations. This constrained what would technically be defined as a Netbook. Microsoft used these same definitions, chief among them the tiny screen size, to offer Windows XP. This had a side-effect of extending the Windows XP lifecycle even more. When we finally celebrated the end of Windows XP it was years longer than originally planned entirely due to Netbooks, though the industry would remember this as a story of the failure of Vista in general.

In the Spring of 2008 after what could be dubbed the Asus Eee PC holiday season, Intel announced the name of the chipset to power Netbooks, ATOM®. With that both Intel and Microsoft were all in on Netbooks, and so were all the OEMs. The collective positioning of Netbooks was as a companion to a primary computer, though that was just marketing. Intel called them MID, for mobile internet device, a third category of device that was neither a mobile phone nor a laptop but a highly portable companion computer. A lot of customers genuinely bought Netbooks as their new laptop.

The industry was filled with concerns over margins from these devices. Intel chipsets that cost around $100 for a laptop were half that for a Netbook. At under $400 there was little room for either margin or innovation. The New York Times wrote of these concerns in 2008, “[D]espite their wariness of these slim machines, Dell and Acer, two of the biggest PC manufacturers, are not about to let the upstarts have this market to themselves." No OEM was going to be left out of what could potentially be a big shift.

Maintaining low prices, especially around $399, posed some problems, mostly the need to cut back on other components and capacities. Intel dropped many of the required aspects of PCs, such as extensible memory, large disk drives, and DVD drives, in an effort to develop the platform that ASUS and others would follow. The PCs would be like a phone—bare bones with all the components essentially built in with little official extensibility. They would have 10-inch or less screens, presumably because of the category but in practice because small screens were super cheap, and, importantly, Intel did not want to see PCs that might compete with profitable, mainstream 13-inch laptops.

The typical specifications of a Netbook were an ATOM processor, 1 or 2GB memory, Wi-Fi, USB ports, SD card reader, a web camera, and 2GB to 4GB of solid-state storage, as opposed to a spinning disk drive. The screen would be an inexpensive LCD running at 1024x600 resolution, with graphics bumping up against the low-end Vista Capable designation. Windows laptops had not yet incorporated solid state storage at all, which made Netbooks rather novel for techies.

For those keeping track, every single one of those specifications was less than the minimum system specifications for the lowest-end Vista PC. Every. Single. One.

That was the rub. These were low-end PCs in every sense. Some of the specs were borderline awful, most notably the screen resolution of 1024x600 was at the limits of what Windows XP would correctly display. Many interactions with Windows would be tricky with so few pixels and much of the Internet and Windows applications would really struggle. By struggle, it was not uncommon to get into a situation where the button that needed to be clicked was off screen and simply unavailable having run out of display area. At times the display just didn’t show enough rows, or the text was too small to read or edit.

HP, Lenovo, Asus, and others released a flurry of devices all with nearly identify specifications. Even though the devices were identical at the processor and chipset and even power adapter, they differed in keyboards, display, and quality of storage used. These small differences were the full-employment act for tech enthusiast web sites that tracked every development in this hot new category.

Personally, I was really into my Netbook. I was already wedded to the 11” form factor having made a switch to use only super portable PCs anyway. The Netbook was tiny, but the low-end nature of the hardware became an ever-present reminder of the need to make Windows 7 work on much less hardware than Vista did. I ran on a Netbook full time for most of the Windows 7 development cycle. The photo in the previous section of me holding up a Lenovo IdeaPad S10 with “I’m a PC” sticker was my very own Lenovo which I even modded myself to 4GB of RAM and a faster solid-state drive. I managed to blog a few hundred thousand words on the tiny keyboard as well. Where it really came in handy was how I constantly used benchmarks and slow operations to annoy the engineering managers. I loved it, but I was a willing subject.

Behind the scenes thought something else was going on. The reality was that the root of the Netbook was not just OLPC and the next billion PC users, rather it was both Intel and Microsoft or Wintel’s inability to transition to a mobile world. The iPhone that released just before the availability of the Eee PC was built with an ARM chip, which technically is a system-on-a-chip, or SoC. ARM chips were what powered the new generation of devices from portable music players to mobile phones to the new iPhone. A SoC packages much of the whole board of a Netbook into a single chip that specifically includes at least the microprocessor and graphics, small and energy-efficient package. The Intel ATOM line was not quite a SoC, though over time it would evolve to be one, at least in name. What made it possible to run Linux and Windows was the combination of compatibility with Intel instructions and the support for PC-style peripherals such as graphics and storage. Microsoft’s version of Windows for the ARM SoC was Windows Mobile, built on the aged Windows code base.

Intel’s entry, or lack thereof, into mobile computing and building a system-on-a-chip components to power mobile phones contributed to the origin story for ATOM. Intel had substantially invested in an attempt to bring the Intel x86 architecture (or IA, as they called it) to mobile phones. Famously, though, Intel ended up not collaborating with Apple on the iPhone, as CEO Paul Otellini felt Apple’s required price was too low and by some accounts desire for control too high. While the initial foray for ATOM was aimed at OLPC, many would claim that Netbooks were simply an effort to make something out of the failed investment for mobile. Since they were fully compatible at the instruction set level with other chipsets, an idea was to build a new laptop—essentially to rescue their failed entry into mobile chipsets. Since these chips drew less battery power, were smaller in size, and less expensive, Intel decided to suggest OEMs create small and inexpensive PCs with them.

In essence, the Netbook arose out of a desire to make use of a low-end chip that originally was meant to compete with ARM and to be used for mobile. While it was compatible with Windows, it was by most accounts inadequate for modern Windows, especially when it came to graphics. The small screen size, while convenient as a demarcation for the Intel chip product line and Windows XP license, was also a necessity because the graphics chipset could not drive a much larger screen.

Nevertheless, Netbooks were flying off the shelves. They were the talk of the industry. Over the next six to 12 months, the sales of these low-end PCs skyrocketed to 40 million units a year—more than 15 percent of all PCs, which was exactly what OEMs loved, especially ASUS that made a big bet. Unfortunately for all involved, the profits were slim across the ecosystem, and worse, the exuberance was truly cannibalizing laptop sales, though not ASUS, which had the new, hot product and a much smaller laptop business. When I visited southeast Asia or Africa, for example, internet cafés had a half dozen netbooks where a single PC once would have been—quantity of endpoints over quality of experience.

Over the course of 2008 leading up to mid-2009, Netbooks remained great sellers. Even if a Netbook was sold with Linux, the demand for Windows was such that pirated Windows XP that was hacked to fit on the available storage became the standard. That seemed to benefit Microsoft in the long run, I suppose. These PCs were seemingly falling from the sky in massive numbers and while the business side was worried about the pricing of Windows and piracy, the product side was happy with something new, finally. There were review websites entirely devoted to the Netbook craze, chronicling the latest developments. In a sense, what was not to love about low-priced computers if people were reasonably happy?

There was little Microsoft could (or should?) do to thwart the momentum, especially because we did not want to lose to Linux. We would have loved nothing more than the likes of HP or Dell to work on Sony VAIO-style PCs that competed with Apple, but that was not to be for a few more years. The PC ecosystem was once again proving that a race to the bottom on pricing and experience was what drove it.

Over time Netbooks expanded in system specifications, OEMs constantly bumping up against the constraints both Intel and Microsoft put in place on what was a legit Netbook. Some added slightly larger screens, making them even slower because extra stress on the under-powered graphics chips. Some added full-size spinning disk drives. Some were able to be upgraded to 4GB of memory. The problem was that under the hood these were still ATOM chips. They were not good PCs. Were they convenient, lightweight, and portable? Yes, but they were slow and had poor graphics. YouTube videos skipped frames and were jittery. Web sites that used Adobe Flash were mostly unusable. Games were too slow. Using Office was marginal at best. Even battery life was limited to three or four hours.

Netbooks, however, played an indisputably positive role in developing Windows 7. They institutionalized a low-end specification that was in-market and broadly deployed. We had to make Windows 7 work reasonably well on them.

Peak Netbook could perhaps have been best described by questions about when Apple would make a Netbook, which Intel would have loved. Such an introduction would have legitimized the category. The mainstream business press just assumed Apple would enter the category because it needed growth, and it was missing out on the hottest new category selling tens of millions of units.

Steve Jobs’s answer to Netbooks, in a uniquely Apple way, was the MacBook Air, which was announced in late January 2008 just after the Eee PC holiday season. Apple’s answer to a $400 Netbook was a $1,799 premium Mac. Classic Apple. The “world’s thinnest notebook,” said Steve Jobs, and Walt Mossberg said it was “impossible to convey in words just how pleasing and surprising this computer feels in the hand.”

It simply wasn’t in Apple’s thinking to release a sub-optimal product. Even the low-priced point products from Apple are fully capable with little compromise. Netbooks violated that core tenet of Apple and so there was no legitimate reason to expect anything other than for Apple to completely ignore this new category the way the industry defined it.

In April 2009, Tim Cook even took to their earnings call to trash talk Netbooks:

When I look at netbooks, I see cramped keyboards, terrible software, junky hardware, very small screens. It’s just not a good consumer experience and not something we would put the Mac brand on. It’s a segment we would not choose to play in. If we find a way to deliver an innovative product that really makes a contribution, we’ll do that. We have some interesting ideas.

At least through 2009, the MacBook Air remained Apple’s answer to small, portable, and even low-priced computing.

There were, however, enough criticisms of the initial MacBook Air to allow PC makers to look the other way or actively market against it by pointing out the lack of ports and extensibility. Instead, the race to the bottom with Netbooks continued. OEMs and Intel would stick with this approach for several years, while Apple refined its new approach to making laptops, eventually even bringing the price down on the MacBook Air. The result was rather crushing. PC elites quickly started running Windows on MacBook Air hardware simply because the hardware was so good, and Apple even provided instructions for how to do that. I can’t even count the number of meetings with OEMs, email threads around Microsoft, and queries from the press I received, asking, “When will there be Windows PCs like the MacBook Air?”

The ecosystem stuck its collective head in the sand all while we rode Netbooks up and then, in a short time, straight back down. In hindsight, the Netbook runup hid the secular decline in PCs that had begun with the recession following the Global Financial Crisis. It would take years to recognize that, and a massive effort for the ecosystem, and Microsoft, to respond to the ever-improving MacBook Air.

Understanding and somehow addressing the relentless force driving the PC ecosystem to produce what most tech enthusiasts see as lesser quality devices (especially relative to Apple) remained one of our key challenges. Too many saw the rise of Netbooks as the answer to products from Apple while not being able or willing to respond directly. There was a fundamental difference between the volume platform appealing to the “masses” and the premium platform appealing to a smaller and more “well-off” segment of the market. That mental model held until the iPhone and MacBook Air. We had reached a tipping point.

With the iPhone, App Store, and MacBook Air Apple was not simply on a roll but the biggest roll of all time from 2006 through 2008.

At the same time, we were deep into building Windows 7 and starting to feel good about the progress. It was time to get out and share our optimism.

On to 094. First Public Windows 7 Demo

Welcome to Chapter XIII! In this chapter we build Windows 7 and bring it to market. We start with all the forces that were shaping up to “disrupt” Microsoft (in the now classic sense) including the launch of the iPhone, cloud computing, consumer internet services, and even the perception of bloat (in Windows this time.) Each of these on their own would be significant, but they were happening all at once, while we were rehabilitating the team, hoping to ship on time for once. To add to the chaos of the moment, these forces appeared during the largest runup of PC sales, breaking 300 million units, followed by the biggest risk to PC sales growth driven by the Global Financial Crisis. A lot was going on competitively, setting the context in which Windows 7 would be built and launched. I thought about competition a great deal, so there is a great deal in this section.

Back to 091. Cleaning Up Longhorn and Vista

The days of competing head-to-head with our own past releases or with vaguely similar products were over. Windows faced outright substitutes, and they seemed to be working.

The Windows 7 team was progressing through engineering milestones M1, M2, and M3 with the energy and momentum increasing along the way, all while computing underwent radical changes at the hardware, platform, and user-experience layers. Everything appeared to be changing at once, just as we were getting our mojo back.

Microsoft’s products and strategy were being disrupted. We just hadn’t, or perhaps couldn’t, come to grips with the reality.

These disruptive forces appeared over the course of developing Windows 7, each one taking a toll on Microsoft’s platform value proposition. Each contributing to a small but growing chorus of changing times that started with the iPhone.

The iPhone was announced in January 2007, six months before Windows 7 Vision Day. The phone didn’t yet have an app store, an app SDK, didn’t run a full desktop browser, lacked push email for Exchange like a Blackberry, and even omitted basic copy and paste. Nobody at the time thought this phone was relevant in a competitive sense to personal computers. Heck, it even required a personal computer for some tasks.

Nobody except Steve Jobs.

We didn’t know the extent the competitive dynamic would shift a year later, creating a true and unforeseen competitive situation, an existential threat, for all of Microsoft.

I attended the iPhone launch event (rushing to it as it was the same week as CES that year) and walked away with a lot to think about for sure. It was easily one of the most spectacular launch events in the history of computing in my lifetime (after Windows 95 and the 1984 Mac launch and its toaster-size computer with a bitmap display that talked). Steve Jobs said one thing that proved to be incredibly important, with long-term implications overlooked by many [emphasis added, my transcription]:

We have been very lucky to have brought a few revolutionary user interfaces to the market in our time. First was the mouse. The second was the click wheel. And now, we’re going to bring multitouch to the market. And each of these revolutionary user interfaces has made possible a revolutionary product—the Mac, the iPod, and now the iPhone. A revolutionary user interface.

We’re going to build on top of that with software. Now, software on mobile phones is like baby software. It’s not so powerful, and today we are going to show you a software breakthrough. Software that’s at least five years ahead of what is on any other phone. Now, how do we do this? Well, we start with a strong foundation.

iPhone runs OS X. [applause]

Now, why, why would we want to run such a sophisticated operating system on a mobile device? Well, because it’s got everything we need. It’s got multitasking. It’s got the best networking. It already knows how to power manage. We’ve been doing this on mobile computers for years. It’s got awesome security. And the right apps.

It’s got everything from Cocoa and graphics and it’s got core animation built in and it’s got audio and video that OS X is famous for. It’s got all the stuff we want. And it’s built right into iPhone. And that has let us create desktop-class applications and networking, right.

Not the crippled stuff that you find on most phones.

This is real, desktop-class applications.

Most reviews mentioned it, but it did not take up nearly as much airtime as the touch screen. In fact, the absence of support for Adobe Flash in the iPhone browser seemed to even undermine this important fact for most. This important fact was the technology underlying the iPhone—the use of the full operating system was a massively strategic, risky, and difficult choice. Using OS X enabled Apple to gradually enable many Mac features over iterative development cycles, knowing that the code already worked. Apple could do this because it had bigger ideas for how it would break compatibility with the Mac and a bold new model for supporting developers to build third-party software. From the very start, the iPhone was destined to be a complete PC, only rebuilt bit by bit with a modern architecture and API. Not only did the iPhone bet on ever-improving mobile bandwidth as many criticized at the time, but it assumed mobile processors and storage would at least reach parity with the personal computer. In other words, from the very start the iPhone had a truck engine. (This reference will make sense in Chapter XIV.)

Windows had been taking the opposite approach which was to base the mobile platform on a nearly decades-old version of Windows, stripped down even, with thoughts though not goals of perhaps of catching up to the current desktop Windows by adding new code over time. The incredible challenges this decision introduced will become readily apparent, but only with the release of a new Windows phone operating system to compete with the iPhone. The diverging paths of Windows for mobile and laptop/desktop had been cast years earlier.

That summer, I lined up at the AT&T store at Pacific Place in downtown Seattle and picked up my black 4GB iPhone. Who needed 8GB on a phone? Some PCs were shipping with 4GB of storage and all of Windows XP. At the time of the launch announcement, I was quite skeptical of the touch keyboard and said so in an internal mail thread, pointing out if touch screens would work then Windows Phone had already tried and mostly failed. I had been a hardcore Blackberry (and later Palm Trēo) user since before general availability in the late 1990s. I was as much a CrackBerry addict as anyone. Some of the many Windows phones had a stylus like a Palm (or Apple Newton), but I never warmed to those and thought handwriting with a stylus was as dumb as Steve Jobs said at the launch. Within a few hours of having the phone (and a browser!), my worldview changed, especially about touch and especially about the evolution of operating systems. Even lacking copy and paste and relying on the slow AT&T 2G mobile network, you had to really try hard to not be impressed.

I remember emailing the co-founder of Blackberry and asking when there would be a full browser and in a long back and forth thread, he tried to convince me of the implications on battery life, lack of capacity on the phone network, and even the lack of utility. While months earlier I might have been sympathetic, I was now staunchly on the other side of this debate.

Inside the company, the iPhone went largely unnoticed outside of small pockets of people, and of course the phone team. Not because it was not a breakthrough product, but because it did not fully connect to our corporate mail service, Microsoft Exchange, as it was configured and permissioned by Microsoft’s IT group. Only those of us running on the testing servers, dogfood, were able to use an iPhone for email, and even then it had no support for our much-loved rights-managed email and Microsoft directory. There was a significant debate over whether maintaining this capability was good for self-hosting and competitive knowledge or bad for supporting competition. It was also about this time that support for using Blackberry was disabled. I put up a huge battle over this only to delay the inevitable. Making it difficult to fully host on competitive products was short-sighted but impossible to stop, even as a senior executive. This was done to “eat our own dogfood” even if the result meant truly innovative and competitive products would only receive cursory use.

While SteveB’s comments around the iPhone launch have become something of an historically bad take as he somewhat mocked the high price, it is crucially important to understand that he as much reflected as simply shared the collective viewpoint of the entire PC industry, and most of the mobile industry. Collectively, nearly every party underestimated the ability for Apple, with no experience in the mobile industry, to deliver a hit product while also resetting the norms of the mobile carrier business model—the same carriers that Steve Jobs described as “orifices” during an interview two years before the world knew about the phone.

The two fundamental assumptions of the PC industry that guided it to nearly 300 million units a year were low prices with consumer choice and a horizontal layering of hardware and software. This business model and technology architecture together enabled the PC. For all the struggles between OEMs, Intel, Microsoft and more, the success was indisputable, especially relative to the Apple model of premium price, limited choice, and a vertically integrated, single supplier.

Before the iPhone, the mobile phone seemed to be coming around to the PC model, with the Windows phone appearing to make progress against Nokia and Blackberry with many PC OEMs using a Windows phone software license to offer low price and choice to customers. Jobs would take this good news and make it appear bad by showing how Windows Phone was fragmented across phone makers, not a single OS. That drove the Windows Phone team crazy, particularly when Jobs presented the market share as a pie chart in 2008 basically rendering Windows Phone as “Other.” The iPhone would literally trounce the traditional OEM model. Choosing between a bunch of crappy devices is no choice at all, even if they all run a single platform. Jobs had long been making this point about PCs too, except now this all seemed to be working.

He was only partially right because in just a short time, Google would repeat the technical and partnership aspects of the Windows model exactly for phones, while upending the economic model and not charging for a software license, opting to further its advertising business model. Of note, the successful OEMs for Android turned out to be an entirely new set of partners, with none of the existing PC OEMs making the generational transition, perhaps in part because most had tried and failed when using Windows mobile software. The combination of iPhone and Android would leave Microsoft in a hopeless middle, with a shrinking partner ecosystem across OEMs and developers. The iPhone was not a broad or instant hit as is well-documented. History does not easily recall the bumpy first year because the following years went so spectacularly well.

Seemingly, at least for the moment, unrelated, there was that money-losing online retailer named Amazon.com led by Jeff Bezos plugging away on what Steve Jobs might have called a hobby or a side project. The company created Amazon Web Services, AWS, in 2002, though few noticed. The product was relaunched about a year before the iPhone with a new set of capabilities and two main new products: EC2 and S3, providing scalable on-demand computing and huge cloud-based storage. For developers, that marked the birth of the cloud era. Amazingly, perhaps, this new competitive front was not Google, which is where nearly all of our platform angst resided.

For Microsoft, this was all starting to sink in. We had competition on the client and at the infrastructure layer. We were getting squeezed on both of our platform businesses, Windows desktops and Windows servers.

Ray Ozzie (ROzzie) was leading a new project, code name Red Dog, with an incredible group of early Windows NT developers including DaveC. As chief software architect, ROzzie was looking to the cloud to reinvent the future of Windows Server, Microsoft’s dominant and extremely profitable business of running enterprise computing in customer-owned and -operated data centers. The running of important business functions in the cloud was still more than a decade away for Microsoft’s enterprise customers, but Amazon’s clarity of vision and the start-ups that gravitated to the approach were incredible. With the cloud, along came Linux as well. Microsoft already needed to catch up, but you had to look very carefully to believe that to be the case, as it was not at all obvious from a customer perspective.

Microsoft had invested a decade in building out Windows Server, a hardware ecosystem, databases, and the entire .NET stack along with a trained sales force. Suddenly AWS running on Linux with entirely new models of compute was the competitor. No salespeople required. To use AWS, developers just signed up online using a credit card and paid for what computing resources they needed as they needed. No hardware to buy. No servers to set up, maintain, and secure. No running out of disk space or network bandwidth. No need to buy more servers in case there’s a rush on the web site. A whole business could run off from laptop with a browser connecting to Amazon’s data centers in the cloud.

It was nearly impossible for many born of the Windows server world, whether at Microsoft or our customers, to fathom this approach. There were endless debates inside the halls and boardroom of Microsoft over whether this was crazy or fantasy. One of the most common debates centered around storing data and either how much bandwidth would be required to move data to or from cloud servers or how expensive cloud storage would forever be. To me these debates seemed rather like the debates over the browser taking over for Office—it is not a matter of if, but rather when everything aligned to make it possible and how different the resulting solutions would look from what we built today. The very existence of gmail seemed to demonstrate the present reality with the most mission critical of all storage intensive workloads, email.

We got so tangled up on how customers would migrate to the cloud, we did not really consider how much larger the market would be starting from nothing and growing cloud native. The irony of us failing to predict this same type of massive upside was almost too much, in hindsight, since that is exactly what gummed up the mainframe computing world and even the character-based (non-GUI) world of MS-DOS as the PC, Windows, and Office platforms took hold.

Additionally, when faced with disruption while substantially new ideas are fast pulling the market and customers in new directions, the previous generation of innovation does not stop. That is what makes it even more difficult for the new approaches to take hold with incumbents. At every step, we thought we would keep adding features in the direction we were going to keep winning. We were winning big. Microsoft’s revenue growth in 2007 and 2008 was 15% and 18% respectively, with 2008 revenue growing past $60 billion.

One example of this competitive pull was our focus on the biggest competitor to Windows Server, VMware. Why? VMware used virtualization to manage Windows Server and in doing so commoditized Windows as one of many alternative operating systems VMware could manage. It had started off as a brilliant invention for developers to isolate code under development on the desktop growing into a new components of enterprise infrastructure. We had been using it ourselves to simplify testing of Windows and even Office for several years already. If Windows Server could enable competitive virtualization, we could thwart the competition from VMware while also solving for the same scenarios Amazon seemed to be addressing with AWS but without Linux.

VMware was acquired (in a complex series of transactions) by a Windows partner, EMC, which soon after acquired a separate company started by Paul Maritz (PaulMa), the former leader of Windows. Paul transitioned to lead VMware as CEO, where he implemented the same enterprise playbook that helped to make Windows Server a success. VMware was rapidly becoming the enterprise standard for a new wave of enterprise datacenter management, which would turn out to be a stop along the way to a future of cloud. This mattered because it impacted the COSD team’s contribution to Windows Server and put the team in between two different versions of the future, being implemented by two different teams at Microsoft, each with serious competition—Red Dog building out one roadmap and Windows Server another. While one view might be that this was a prudent strategy, another view was that it was a strategy guaranteed to slow our progress to an inevitable future.

Our remaining competitive challenge was faced by Live Services. Competition was constant and coming from all directions. Many competitors came and went quickly, as was the norm in consumer, ad-supported services. Switching costs were low and whole populations changed quickly—it appeared to be a hit-driven business, which was not something Microsoft was geared up to navigate, especially after a failed decade of trying to make consumer and home software a thing with CDROM titles and later web sites. MSN Messenger and Hotmail had hundreds of millions of users, but daily active users (DAUs) were declining and engagement (usage time) overall was dropping. There was a good deal of advertising revenue, hundreds of millions of dollars, but it depended on intrusive display ads that riled users, even though the services were free.

Gmail rapidly became the new de facto leader in “free” email, offering essentially “unlimited” email storage in gigabytes while Hotmail was trying to build a business charging for extra megabytes. Originally announced as an April Fool’s prank in 2004 and released in beta, by the time the Windows 7 vision meeting took place Gmail had finally removed the exclusive invitation signup requirement. Though remained in beta, the service was exploding in use. While it would be a few years until Gmail surpassed Hotmail, Hotmail almost immediately stopped seeing growth and started to see a decline in engagement. Gmail was not a gimmick; under the hood was an enormously innovative database and operational capability. Gmail had no advertisements. None.

MSN Messenger, eventually Live Messenger, had become enormously popular around the world, especially outside the United States with hundreds of millions of active users. It too was facing an existential competitive threat. This time from Skype, a Swedish, Danish, and Estonian invention that offered free voice calls from almost any platform, notably both PC and Mac. While Messenger was often used to arbitrage SMS fees, Skype was arbitraging voice and creating a movement that would permit much of the world to skip using landlines for overseas calls when mobile minutes were incredibly expensive. Video calling was introduced as well, and while Messenger already had this capability, the cross-platform nature of Skype, as well as the focus on voice connections to local land lines, made for a much more compelling offering. Microsoft would finally acquire Skype from eBay (who had acquired it in 2005) in 2011 when it reached almost 300 million users worldwide, more than Messenger had achieved. In 2007-2009, Windows Live was still competing with Apple, Google, Yahoo, Skype, MySpace, and a host of category leaders across photos, blogging, and video. That was a lot.

SteveB and Kevin Johnson spent a great deal of time and energy on the potential of acquiring Yahoo, the dominant leader with which MSN competed. Such a deal would have added to my challenges given their email and messenger services were suffering much the same way. We might have gained in search and content services, but we would have added productivity services that were also losing share just as Windows Live seemed to be.

Apple struggled to find its way through the cloud and services world, even with the launch of iPhone. Apple’s decidedly client and device focused approach was quite similar to how we saw Live Services and Windows evolving together. The services would be augmented by rich “desktop class” applications for photos, video, messaging, blogging, and even productivity. Apple for years had been selling a suite of creativity products, iLife, later adding new productivity tools, notably Keynote for slide shows, called iWork. A collection of web services was originally called iTools then later rebranded as the .Mac service (pronounced “dot Mac”) and included email, online storage, and backup. In the summer of 2008, the service was rebranded MobileMe in a very bumpy launch that was not widely praised. After eight years and a good deal of iteration, Apple continued to work to find its way even as iPhone success grew.

The most disruptive announcement from Apple came a year after the initial iPhone launch. In 2008, Apple announced and then brought to market a software development kit with APIs to build third-party apps for the iPhone. This also included an entirely new store for software distribution and economic model for developers, the App Store. It is almost impossible to overstate the leap the App Store brought to computing. The PC was drowning in viruses and malware because of the internet—the ability in a few clicks to download software seemed wonderful at first, but then quickly became a cesspool of awful software that at best simply degraded the PC experience. Additionally, the world of PC software had stagnated simply because it was so large that it became almost impossible for a new desktop product to gain awareness, distribution, and enough sales to support a pure-play software company. In fact, Skype might be the most innovative native, though cross-platform, application to break through outside of browsers, and we acquired it in 2011 before it was profitable. While some would view the App Store as a sort-of closed ecosystem, it was literally the solution to the problems plaguing the PC. The Apple bet on OS X meant that there was a robust and proven platform and toolset to serve as a foundation, plus software distribution and economics. Developers hardly had all of OS X, but they definitely had a lot of it and Apple could add more over time.

Microsoft was steeped in a competitive mindset across every generation of leader and from many perspectives. BillG never missed an opportunity to cite some positive attribute or significant asset of a competitor. SteveB brought his relentless sense of competition from childhood math camp and sports. MikeMap and JeffH instilled this in all of us back on my first team with an intense business rigor. In working with the Windows and Windows Live role, I was faced with competition from more directions and with more depth than I ever thought possible. The team, at first as I moved into the role, seemed to consistently minimize this new reality. It would be rude to say Microsoft was in denial. I don’t think it would be unfair to say that after years of winning and even feeling like we had beaten back Linux and open source, Microsoft had become much more focused on its own universe of customers and their problems which was mostly immune to influences from outside that gravitational sphere.

In hindsight, that was what happened when two factors combined to create a step-function change in product trajectory.

First, the existence of a single massive product, Windows and the Intel PC ecosystem, Wintel, created constraints for those looking to build entirely innovative products. While many built products thinking of Wintel as an ingredient, amplifying the platform without posing a risk, a small percentage of risk-takers saw a different world. They saw the shortcomings of Wintel as an opportunity to reinvent. They built as though the leaders were not there and built what, in their eyes, the world should look like, whether they achieved critical mass success or not. Each of these new competitors had a worldview that revisited underlying assumptions about Windows and Windows Live and the PC ecosystem. Competitors assumed any web browser as the user interface, connected over the internet to Linux servers running open source software—no Windows Server or .NET at all. Google, Facebook, and a constellation of start-ups in Silicon Valley embraced this model as though Microsoft never existed. Even when it came to Microsoft Office, most new companies in Silicon Valley operated as though it was an insignificant part of the software landscape. In 2008, while Windows 7 was in testing as we tried to bring Internet Explorer back from hibernation, Google released the Google Chrome browser and with it putting an end to even that sliver of Microsoft as part of the next wave of innovations.

Second, the incumbent leader had to mess up. Customers generally didn’t spontaneously change, even if there was something better to switch to, because of processes, habits, and costs, and they don’t change all at once. Leaders messed up by ignoring new technologies, especially as over time little technologies added up to something material. Also, a risk was a failure to execute and deliver new products to market, simply dropping the ball. Microsoft mired in the journey through Windows Longhorn and a executing a Windows strategy put forth in the mid-1990s had indeed dropped the ball. It was increasingly difficult to appreciate or even see changes to the technology landscape when the company’s decision-making context is so dominated by goals, challenges, and issues entirely of its own making.

Waiting to pick up the ball was the competition born of the internet and web. While many wished to connect this potential disruption of Microsoft to the antitrust events and resulting settlement in the early 2000s, none of Apple’s iPhone, Amazon’s AWS, or Google’s Search or Gmail had anything to do with the trial and resulting settlement. Where some like to claim Microsoft was distracted, they would be wrong. If Microsoft was distracted, it was by simply trying to finish Windows Vista or compete with VMware or IBM or even Linux, or executing on our own plans and growth and dealing with issues like software security and quality. This wasn’t about the browser, the price of Windows, or even what software was installed on a PC. Those had already become old battles and irrelevant to the rapid structural changes that happened in software (the kind that produced Microsoft, then Windows, then Office in the first place). As fast as one company can rise to success, another can do the same in equally unexpected or counterintuitive ways. This was the argument or perhaps the defense Microsoft and BillG had put forth time and time again. No matter how many times BillG said this to the press, customers, or regulators none would believe him…even as it was taking place.

This was all happening as I was trying to get our house in order and make progress. Things just weren’t as simple as they were in 2005 when OEMs were just waiting on a new release of Windows. The PC makers were looking particularly unhealthy and deeply concerned about the rise of mobile phones on top of normal concerns about the price of components, delays in Windows, and the chip roadmap from Intel—would phones be an opportunity for PC OEMs or would they prove to be a generational change in hardware leaders as well? Should they be considering Linux on the desktop as a replacement of Windows given its popularity on servers?

SteveB, and increasingly the Board, had a lot of questions and concerns surrounding the competitive dynamic. Some board members were close to the hardware ecosystem and would oscillate between certainty that the ecosystem would deliver and that it would not. International board members were using Skype to talk to their families. Everyone wanted to know how to connect from their iPhones to their Microsoft corporate email or Windows Live personal mail, or why iTunes was so slow on Vista, or why Mac Excel was so different from Windows Excel as new Mac owners were discovering.

In fact, they were all asking how long it would be before they did everything on mobile phones. There was also deep concern over browsers, knowing it had been five years since the last release of Internet Explorer in 2001 and work had all but stopped. All of them wanted a PC as cool as a MacBook. Microsoft board members had a budget for PC purchases and always wanted to know the most Apple-competitive Windows laptop to buy, and for much of the duration of Windows 7 we had no answers. Each time I attended a Board meeting, I had to respond to all of these questions again.

Like Captain Kirk in ST:WOK (Star Trek: Wrath of Khan), I would look around and think, “Please, please…give us time…the bridge is smashed, computers inoperative.” We were rebuilding the team and trust with customers and OEMs. Windows 7 was going to pave the way for us to do big new things. There was little more we could do than get that done.

Whether it was desperation, a lack of alternatives, or simply misplaced confidence in the team, the questions kept coming yet there were few questions about Windows 7. Many were already looking beyond Windows 7, thinking, and plotting as though delivering Windows 7 was some sort of no-brainer.

Looking back, it was equally easy to ponder the radical idea of basically skipping Windows 7 and going straight for something to compete on these many fronts. We could theoretically catch up to these multiple competitive forces and not miss an entire cycle of innovation if everything was aimed at mobile and cloud. One big mega-strategy to build a new Windows, a new phone, and integrated cloud services.

That would have been absurd. It was the opposite of possible. The team was still recovering from the Vista release with its portfolio of stretch goals, to put it kindly, that did not go as planned. The last thing we should have even come up with was some sort of Longhorn-redo. Frankly, what we planned for Windows 7 was kind of crazy, given the recent track record. We planned Windows 7 and all the features with the assumption the team could deliver a major update to Windows, on time. Additionally, the Server team and its customers remained not only unconvinced of the cloud but actively campaigned against it. Besides, we had RedDog. Windows 7 had to serve both of those.

The impact and the constraints of the past had long-lasting effects.

While the Board was anxious about the post-Vista landscape, the technical trade press, mostly made up of Microsoft watchers, remained tuned to Windows. It was the product with the most familiarity, and the Vista release was causing difficulties. But the double-edged sword of the beat reporters was that they covered what we were up to, but not so much what we should have been up to, until it was too late. It would still be a few years before most reporters made the switch to Mac, but the switch to iPhone was happening quickly bringing renewed attention to Apple laptops.

At the same time, any little thing we did was chum in the water for the dozens of beat reporters covering Microsoft.

In the summer of 2007 word leaked out about the impending service pack for Windows Vista previously described. It was not a horrible leak, but it scrambled our OEM partners and immediately froze the few Vista enterprise deployments in process. The field sales team was livid that they had not been briefed on the release, again arguing for more transparency from Redmond. The problem was we had just gone through five years of transparency and every constituent was annoyed, at best. I wrote a 3000-word internal blog post “Transparency and disclosure” where I tried to put forth the idea that being transparent isn’t compatible with being good partners, but we needed to aspire to translucency (excerpt follows).

Transparent. Easily seen through or detected; obvious.

Translucent. easily understandable; lucid.

One topic I have been having an interesting time following has been the blogs and reports that speculate about how Windows will go from being an open or transparent product development team to being one that is “silent” or “locked down”. Much of this commentary seems to center around me personally, which is fine, but talks about how there is a Sinofsky-moratorium on disclosure. I think that means I owe it to the team to provide a view on what I do mean personally (and what I personally mean to do), of course I do so knowing that just by writing this down I run this risk of this leaking and then we’ll have a round of phone calls and PR management to do just with regards to “Sinofsky’s internal memo on disclosure”. But I thought it would be worth a try.

…

Customers and partners want to know about SP1 for Vista. Actually they need to know. We want to tell them. But we want to do so when our plans and execution allow that communication to be relatively definitive. We are not there yet. So telling folks and then changing the plans causes many more challenges than readily apparent. While it might sound good on paper to be “transparent” and to give a wide open date range and a wide open list of release contents, we all know that these conversations with customers don’t end with the “we’ll ship by date and we’ve prioritized ”. Folks do want to know “did you fix this bug?” That is reasonable, but we don’t have all those answers and thus we cannot have a reasonably consistent and reliable communication…yet. We are working towards that. While there is clearly a challenge in the near term in not offering details, this challenge is much less than if we get the wrong information out there and we have to reset and unset expectations. Even among our enterprise customers, for whom this type of information is routine, we have a long history of really scrambling these most valuable customers with “information” that turned out to be “misinformation”. The difference we are trying to highlight is the difference between transparency in what we’re “thinking” and transparency in what we’re “doing”. Everyone wants to know what we’re thinking, but making it clear that those are thoughts versus “doing” is a subtlety lost on a mass audience.

…

I know many folks think that this type of corporate “clamp down” on disclosure is “old school” and that in the age of corporate transparency we should be open all the time. Corporations are not really transparent. Corporations are translucent. All organizations have things that are visible and things that are not. Saying we want to be transparent overstates what we should or can do practically—we will share our plans in a thoughtful and constructive manner.

This too leaked. It became known as the “Sinofsky omerta” (awful) and the idea of being translucent was always said snidely. We had gone nine-plus months without any substantial forward-looking discussion of Windows. The reporters covering Microsoft were restless. The leaked blog post only served to amplify any other leaks.

A college student-blogger in Australia, Long Zheng, who had become somewhat of a canary in tracking Windows evolution, happened to catch an October 2007 college recruiting talk at the University of Illinois given by Eric Traut (EricTr), a Distinguished Engineer and one of the senior architects in COSD and a core member of the Windows NT architecture team. Eric was a key inventor of the hardware emulation technology used by Apple as it transitioned microprocessors the first time in the 1990s and then an early pioneer of virtualization at a competitor to VMware, Connectix, that Microsoft acquired in 2003.

EricTr presented a wonderfully detailed talk on the role of virtualization in modern computing, describing the work he did along with a number of the most senior people on Windows. He described the architecture of a modern scalable OS and the evolution of Windows over time. He even showed some code.

To that blogger, he had just seen a demo of Windows 7 or at least exactly what Windows 7 should be.

Zheng saw the presentation on a video posted on Microsoft’s own Channel9 website and promptly wrote a blog post about the future of Windows 7, referring to the talk as a “demonstration of Windows 7.”

It wasn’t that at all. What was so exciting, though?

Over the years there was that constant murmur underlying the evolution of Windows and its expression of bloat—expansion of code size, decreasing performance, the requirement that PCs have more and more memory just to keep up. Everyone’s PC got slower over time because of bloat. EricTr’s demonstration inadvertently played into this narrative and provided a huge hope for improvements in the future.

To implement virtualization, the OS kernel was being worked on to further reduce the impact it would have on memory and CPU consumption. Eric demonstrated this new “minimal kernel,” which he dubbed MinWin. This would be exactly the solution to every problem with PCs—if Windows was good but a bit bloated, then certainly a minimum Windows, or MinWin, would be what everyone would want. It would do everything Windows did but with minimal code. Who would not want that? If one set out to create an ideal branding to describe a release of Windows for every tech enthusiast, MinWin was it. As I read the blogger’s account, I thought “Oh my gosh, this is Office Lite, but for Windows”—it was a hypothetical product that did everything it needed to do, but was easier, smaller, faster, and lighter. What’s not to love? (See 076. Betting Big to Fend Off Commoditization)

Eric’s talk was about virtualization and scaling Windows Server. It was not at all about Windows the way most people thought about it. Eric described what he was showing as follows:

We created what we call MinWin. Now, this is an internal only, you won't see us productizing this, but you could imagine this being used as the basis for products in the future. This is the windows 7 source code base, and it's about 25 megs on disk. You compare that to the four gigs on disk that the full windows Vista takes up.

Everyone on the team, most certainly a COSD architect, knew the bloggers description that this was Windows 7 was incorrect—Eric even said so. Even assuming Windows was bloated, the OS kernel wasn’t the culprit. We ended up spending a lot of time with the press figuring out how to reduce the expectations and to not consider the next Windows some sort of “new kernel.” This backfired because they heard us downplaying expectations for the release overall. With our silence, this made sense, given the Windows history of overpromise and underdeliver. As word spread, we began to hear from enterprise customers who thought that a new kernel would introduce compatibility concerns, especially in the server. This was neither a practical nor theoretical issue.

The challenge this created was indirectly related to bloat. Rather, it created a perception that Windows 7 was going to be substantially “less bloated” (whatever that meant) than Windows Vista. That prompted people to talk about comparisons to the now, post-Vista, much-loved Windows XP. Nothing in the product plan had changed, but there was suddenly a perception Windows 7 would be dramatically improved.

The public expectations for the release went up, as if they weren’t already sky high. When the outer reaches of the company saw these stories, particularly where there were direct connections to customers, the optimism was contagious.

Scary? Sure. Except we already had thousands of people working on that very opportunity—making a leaner, more efficient Windows while also making it do a lot of new stuff.

The absence of information gave people ample room to create their own worldview. Many, especially SteveB and the OEM team, wanted to use this, or one of the other rumor cycles, to get out there with more information. The team was still in the early stages of executing, and we needed to make more progress. I did not want to be out there just yet. Besides, all I could do at that moment would be counter the now exaggerated expectations which would create more confusion and likely cause another news cycle condemning Windows 7 to be a limited or incremental update to Vista.

Just as the Fall computer selling season was gearing up, a new type of computer was all the rage in Asia and about to hit the US market. Maybe these “netbook computers” would breathe some life into Vista and buy us some time before we started talking about Windows 7?

On to 093. Netbook Mania

Whenever you take on a new role you hope that you can just move forward and start work on what comes next without looking back. No job transition is really like that. In my case, even though I had spent six months “transitioning” while Windows Vista went from beta to release, and then even went to Brazil to launch Windows Vista, my brain was firmly in Windows 7. I wanted to spend little, really no, time on Windows Vista. That wasn’t entirely possible because parts of our team would be producing security and bug fixes at a high rate and continuing to work with OEMs on getting Vista to market. Then, as was inevitable, I was forced to confront the ghosts of Windows Vista and even Longhorn. In particular, there was a key aspect of Windows Vista that was heavily marketed but had no product plan and there was a tail of Longhorn technologies that needed to be brought to resolution.

Back to 090. I’m a Mac

Early in my tenure, I received an escalation (!) to “fund” Windows Ultimate Extras. I had never funded anything before via a request to fund so this itself was new, and as for the escalation. . . I had only a vague idea what Ultimate Extras were, even though I had recently returned from the Windows launch event in Brazil where I was tasked with emphasizing them as part of the rollout. The request was deemed urgent by marketing, so I met with the team, even though in my head Vista was in the rearview mirror and I had transitioned to making sure servicing the release was on track, not finishing the product.

The Windows Vista Ultimate SKU was the highest priced version of Windows, aimed primarily at Windows enthusiasts and hobbyists because it had all the features of Vista, those for consumers, business, and enterprise. The idea of Ultimate Extras was to “deliver additional features for Vista via downloadable updates over time.” At launch, these were explained to customers as “cutting-edge features” and “innovative services.” The tech enthusiasts who opted for Ultimate, for a bunch of features that they probably wouldn’t need as individuals, would be rewarded with these extra features over time. The idea was like the Windows 95 Plus! product, but that was an add-on product available at retail with Windows 95.

There was a problem, though, as I would learn. There was no product plan and no development team. The Extras didn’t exist. There was an Ultimate Extras PUM, but the efforts were to be funded by using cash or somehow finding or conjuring code. This team had gotten ahead of itself. No one seemed to be aware of this and the Extras PUM didn’t seem to think this was an issue.

As the new person, this problem terrified me. We shipped and charged for the product. To my eyes the promise, or obligation if one prefers, seemed unbounded. These were in theory our favorite customers.

The team presented what amounted to a brainstorm of what they could potentially do. There were ideas for games, utilities, and so on. None of them sounded bad, but none of them sounded particularly Ultimate, and worse: None existed.

We had our first crisis. Even though this was a Vista problem, once the product released everything became my problem.

The challenge with simply finding code from somewhere, such as a vendor, licensing from a third party, or something laying around Microsoft (like from Microsoft Research), was that the journey from that code to something shippable to the tens of millions of customers running on a million different PC configurations, in 100 languages around the world, and also accessible, safe, and secure, was months of work. The more content rich the product was, in graphics or words, the longer and more difficult the process would be. I don’t know how many times this lesson needed to be learned at Microsoft but suffice it to say small teams trying to make a big difference learned it almost constantly.

And then there was the issue of doing it well. Not much of what was brainstormed at the earliest stages of this process was overly compelling.

With nothing particularly ultimate in the wings, we were poised for failure. It was a disaster.

We set out to minimize the damage to the Windows reputation and preserve the software experience on PCs. Over the following months we worked to define what would meet a reasonable bar for completing the obligation, unfortunately I mean this legally as that was clearly the best we could do. It was painful, but the prospect of spinning up new product development meant there was no chance of delivering for at least another year. The press and the big Windows fans were unrelenting in reminding me of the Extras at every encounter. If Twitter was a thing back then, every tweet of mine would have had a reply “and…now do Ultimate Extras.”

Ultimately (ha), we delivered some games, video desktops, sound schemes, and, importantly, the enterprise features of Bitlocker and language packs (the ability to run Vista in more than one language, which was a typical business feature). It was very messy. It became a symbol of a lack of a plan as well as the myth of finding and shipping code opportunistically.

Vista continued to require more management effort on my part.

In the spring of 2007 shortly after availability, a lawsuit was filed. The complaint involved the little stickers that read Windows Vista Capable placed on Windows XP computers that manufacturers were certifying (with Microsoft’s support) for upgrade to Windows Vista when it became available. This was meant to mitigate to some degree the fact that Vista missed the back to school and holiday selling seasons by assuring customers their new PC would run the much publicized Vista upgrade. The sticker on the PC only indicated it could run Windows Vista, not whether the PC also had the advanced graphics capabilities to support the new Vista user experience, Aero Glass, which was available only on Windows Vista Home Premium. It also got to the issue of whether supporting those features was a requirement or simply better if a customer had what was then a premium PC. The question was if this was confusing or too complex for customers to understand relative to buying a new PC that supported all the features of Vista.

A slew of email exhibits released in 2007 and 2008 showed the chaos and tension over the issue, especially between engineering, marketing, sales, lawyers, and the OEMs. One could imagine how each party had a different view of the meaning of the words and system requirements. I sent an email diligently describing the confusion, which became an exhibit in the case along with emails from most every exec and even former President and board member Jon Shirley (JonS) detailing their personal confusion.

The Vista Capable challenge was rooted in the type of ecosystem work we needed to get right. Intel had fallen behind on graphics capabilities while at the same time wanted to use differing graphics as part of their price waterfall. Astute technical readers would also note that Intel’s challenge with graphics was rooted deeply in their failure to achieve critical mass for mobile and the resulting attempt to repurpose their failed mobile chipsets for low-end PCs. PC makers working to have PCs available at different price points loved the idea of hardline differentiation in Windows, though they did not like the idea having to label PCs as basic, hence the XP PCs were labeled “capable.” Also worth noting was that few Windows XP PCs, especially laptops, were capable of the Home Premium experience due to the lack of graphics capabilities. When Vista released, new PCs would have stickers stating they ran Windows Vista or Windows Vista Basic, at least clarifying the single sticker that was placed on eligible Windows XP computers.

Eventually, the suit achieved class-action status, always a problem for a big company. The fact that much of the chaos ensued at the close of a hectic product cycle only contributed to this failure. My job was to support those on the team that had been part of the dialogue across PC makers, hardware makers, and the numerous marketing and sales teams internally.

The class-action status was eventually reversed, and the suit(s) reached a mutually agreeable conclusion, as they say. Still, it was a great lesson in the need to repair both the relationships and the communication of product plans with the hardware partners, not to mention to be more careful about system requirements and how features are used across Windows editions.

In addition to these examples of external issues the Vista team got ahead of itself regarding issues related to code sharing and platform capabilities that spanned multiple groups in Microsoft.

The first of these was one of the most loved modern Windows products built on top of Windows XP, Windows Media Center Edition (WMC, or sometimes MCE). In order to tap into the enthusiasm for the PC in the home and the convergence of television and PCs, long before smartphones, YouTube, Netflix, or even streaming, the Windows team created a separate product unit (rather than an integrated team) that would pioneer a new user interface, known as the 10-foot experience, and a new “shell” (always about a shell!) designed around using a PC with a remote control to show live television, home DVD discs, videos, and photos on a big screen, and also play music. This coincided with the rise in home theaters, large inexpensive disk drives capable of storing a substantial amount of video, camcorders and digital cameras, and home broadband internet connections. The product was released in 2002 and soon developed a relatively small but cultlike following. It even spawned its own online community called “Green Button,” named after the green button on the dedicated remote control that powered the shell’s 10-foot user interface.

The product was initially sold only with select PCs because of the need for specific hardware capabilities. Later, with Windows Vista (and Windows 7), WMC was included in the premium editions. The usage based on both sales and the telemetry collected anonymously was low and the repeat usage was a small fraction of even those customers. Nevertheless, there were vocal fans, and we had no plans to give up.

WMC was hitting real challenges in the market, though, especially in the United States, where television was moving from analog CATV to digital, and with digital came required set-top boxes and a new and not quite available technology called CableCARD, required to decrypt the cable signal. Not only did this make things difficult for WMC, but it made things difficult for anyone wanting to view cable tv, as if the encrypted analog channels were not difficult enough already. Everyone trying to use CableCARD had a story of trying to activate the card that included essentially a debug interface, typing in long hex strings, awaiting a “ping” back from the mysterious “head end.” The future for the innovative TV experience in WMC was looking bleak.

Additionally, WMC was bumping up against the desires of the Xbox team to expand beyond gaming. The Xbox team had recently unveiled a partnership with the new Netflix streaming service to make it available on Xbox. Some of the key leaders on WMC had moved from Windows to the Xbox organization and began to ask about moving the WMC team over with them.

At the time, I was up to my elbows in looking at headcount and orgs and was more than happy to move teams out of Windows, especially if it was straightforward, meaning they could just pick up the work and there was no shipping code being left behind unstaffed. This quickly became the first debate, in my entire time at Microsoft, over headcount and budgets because the destination organization was under tight revenue and expense review.

The WMC team was, surprisingly, hundreds of people, but it also had dependencies on numerous other teams across networking, graphics, and the core user experience. We could easily move the core WMC team but getting a version of WMC integrated with the new engineering system and to-be-developed delivery schedule (which we were planning) was a concern. Of course the team wanted to move to Xbox but had little interest in delivering WMC back to Windows, especially as the overall engineering process changed. They literally thought we would just move all the headcount and team and then create a new WMC team. They had awful visions of being a component on the hook to meet a schedule that was unappealing. We could not just give up on WMC, even with such low usage, without some sort of plan.

I learned that moving humans and associated budgets was fine. But CollJ had been working with finance on headcount and was told we had to also move revenue, something I had never heard of. I had no idea how that might even work. In Vista, MediaCenter was part of the Home Premium and Ultimate SKU, and no longer a separate product like it was for Windows XP. How could one arrive at the revenue for a single feature that was part of a SKU of Vista? Perhaps back when WMC was a separate product this made sense, but at the time it seemed like an accounting farce. In fact, the judge in the Vista Capable lawsuit even removed class action status because of an inability to determine which customers bought Windows Vista because of which premium features.

Microsoft had been divided into seven externally reported business segments; each quarterly earnings filing with the SEC reported a segment as though it were an independent business. The result of this was more visibility for the financial community, which was great. Internally these segments did not line up with the emphasis on code-sharing, feature bundling, or shared sales efforts. For example, from a product development perspective there was a lot of code sharing across all products—this was a huge part of how Microsoft developed software. Costs for each segment could never accurately reflect the R&D costs. An obvious example was how much of Windows development could/would/should be counted in the Server and Tools segment, given that so much of the server product was developed in the Windows segment.

My view was that there was a $0 revenue allocation for any specific feature of Windows—that was the definition of product-market fit and the reality that nobody bought a large software product for any single feature. This was always our logic in Office, even when we had different SKUs that clearly had whole other products to justify the upsell. Office Professional for years cost about $100 more than Office Standard simply for the addition of the Access database. We never kidded ourselves that Access was genuinely worth billions of dollars on its own.

Over several weeks, however, we had to arrive at some arbitrary number that satisfied finance, accounting, tax, and regulatory people. To do this, we moved the hundreds of people working on WMC to XBox along with a significant amount of Windows unit revenue per theoretical WMC customer. It wasn’t my math, but it added up to a big number moving to the Entertainment segment. In exchange, we hoped we would get back a WMC that had improved enough to satisfy an enthusiastic fanbase, knowing that the focus was shifting to Xbox.

Given the fanbase externally, including several of the most influential Windows writers in the press, there was a good chance that such a seemingly arbitrary organization move would leak. The move would be viewed by many as abandoning WMC. I used an internal blog post to smooth things over, describing my own elaborate home audio/video system, which included Windows Vista Ultimate and pretty much every Microsoft product. In doing this, I chalked up a ton of points from tech enthusiast employees, showing that I had a knee-deep knowledge of our products—something most assumed an Office person wouldn’t have, perhaps like BillG not understanding why a C++ developer knew Excel 15 years earlier when I first interviewed to work for him. In practice, my post made it clear that keeping this technology working to watch TV casually at home was impossible. I hated my home setup. It was ridiculous.

As part of the final Longhorn cleanup, I also needed to reconcile the strategic conflicts between the Windows platform and the Developer Tools platform, and as the new person I found myself in the middle. It was a battle that nobody wanted to enter for fear of the implications of making the wrong technology bet.

The Developer division created the .NET platform starting in the early 2000s to build internet-scale, web server applications delivered through the browser primarily to compete with Java on the server. It was excellent and remains loved, differentiated, and embraced by the corporate world today. It crushed Java in the enterprise market. It was almost entirely responsible for the success Windows Server saw in the enterprise web and web application market.

The .NET client (desktop programs one would use on a laptop) programming model was built “on top” of the Windows programming model, Win32, with little coordination or integration with the operating system or the .NET server platform. This created a level of architectural and functional complexity along with application performance and reliability problems resulting in a messy message to developers. Should developers build Win32 apps or should they build .NET apps? While this should not have been an either/or, it ended up as such because of the differing results. Developers wanted the easier-to-use tools and language of .NET, but they wanted the performance and integration with Windows that came from Win32/C++. This was a tastes great, less filling challenge for the Developer division and Windows. In today’s environment, there are elements of this on the Apple platforms when it comes to SwiftUI versus UIKit as searching for that debate will find countless blog posts on all sides.

It was also a classic strategic mess created when a company puts forth a choice for customers without being prescriptive (or complete). Given a choice between A and B, many customers naturally craft the third choice, which is the best attributes of each, as though the code was a buffet. Still other customers polarize and claim the new way is vastly superior or claim the old way is the only true way and until the new way is a complete superset they will not switch. Typical.

Longhorn aimed to reinvent the Win32 APIs, but with a six-year gap from Windows XP it was filled by the above .NET strategy when it came to Microsoft platform zealots. The rest of the much larger world was focused on HTML and JavaScript. At the same time, nearly all commercial desktop software remained Win32, but the number of new commercial desktop products coming to market was diminishingly small and shrinking. Win32 was on already life support.

The three pillars of Longhorn, WinFS, Avalon, and Indigo, failed to make enough progress to be included in Vista (together these three technologies were referred to as WinFX.) With Vista shipping, each of these technologies found new homes, or were shut down. I had to do this last bit of Vista cleanup which lingered long after the product was out the door.

WinFS receded into the world of databases where it came from. As discussed, it was decidedly the wrong approach and would not resurface in any way. Indigo was absorbed into the .NET framework where it mostly came from. Avalon, renamed Windows Presentation Foundation (WPF), remained in the Windows Client team, which meant I inherited it.

WPF had an all-expansive vision that included a unique language (known as XAML) and a set of libraries for building graphical and data-rich applications. Taken together and to their logical end point, the team spoke of Avalon as being a replacement for HTML in the browser and also .NET on the client. This was the reason work had stopped on Internet Explorer, as the overall Longhorn vision was to bring this new level of richness to browsing. From the outside (where I was in Office) it seemed outlandish, and many on the outside agreed particularly those driving browser standards forward. Still, this opened a second front in the race to improve Win32, in addition to .NET. All along WPF would claim allegiance and synergy with .NET but the connections were thin, especially as WPF split into a cross-platform, in-browser runtime and a framework largely overlapping with .NET. When I would reflect on WPF I would have flashbacks to my very first project, AFX, and how we had created a system that was perfectly consistent and well-architected yet unrelated to everything, in addition to being poorly executed.

But what should developers have used—classic Win32, or the new frameworks of .NET, WPF, or something else I just learned about called Jolt? These were big decisions for third-party developers because there was an enormous learning curve. They had to choose one path and go with it because the technologies all overlapped. That was why I often called these frameworks “fat frameworks,” because they were big, slow, and took up all the conceptual room. Importantly, the common thread among the frameworks was their lack of coherence with Win32—they were built on top of Win32, duplicating operating system capabilities. Fat was almost always an insult in software.

The approach taken meant the idea of both .NET on the client and WPF were built on the shakiest of foundations, and what we were seeing with .NET on the client was as predicted based on all past experiences. I had a very long and ongoing email thread (which I later turned into an Office Hours blog post) with the wonderfully kind head of Developers, S. Somasegar (Somase, aka Soma), on this topic where I pushed and pushed on the challenges of having fat frameworks emerge while we have a desire to keep Windows relevant. As I wrote the email, I was struck by how similar this experience was to the one I had twenty years earlier as we built, and then discarded, the fat framework called AFX. While history does not repeat itself, it does rhyme. Many would debate the details, but fundamentally the team took the opposite path that we did years earlier. I recognize writing this today that some are still in the polarized camp and even today remain committed to some of these technologies. One of the most successful tools Microsoft ever released was Visual Basic, which was an enormously productive combination of framework, programming language, and development tool. That was not what we had here.

Nothing with developers and Microsoft was ever simple. Microsoft’s success with developers was often quite messy on the ground.

Along with large product development teams, there was also a large evangelism team responsible for gaining support or traction of new technologies with developers. This team was one of the gems of Microsoft. It masterfully built a community around developer initiatives. It was largely responsible for taking the excellent work in .NET for Server and landing it with thousands of enterprise developers. In fact, part of that challenge was that the evangelism team had moved to the Developer division where the org chart spoke more loudly than the overall company mission and the priority and resourcing of evangelism tilted heavily toward .NET in all forms over the declining Win32 platform.

As a counterexample, the evangelism team seeing the incoherence in product execution provided significant impetus to force the Windows NT product to fully adopt the Windows API paving the way for Microsoft’s Win32 success. I previously shared my early days story as a member of the C++ team pointing out how Windows NT differed from classic Windows APIs in Chapter II, I Shipped, Therefore I am. Many of the lessons in how divergent the two Windows APIs were surfaced in the well-known Windows NT porting lab run by the evangelism team, where developers from around the world would camp out for a few weeks of technical support in moving applications to Windows NT or Windows 95, before both shipped.

Perhaps an organization-driven result was a reasonable tradeoff, but it was never explicit, as was often the case with cross-divisional goals. In many ways the challenges were accelerated by our own actions, but without ever making that a clear goal we would spend too much time in meetings dancing around the results we were seeing in the erosion of Win32.

The evangelism team didn’t ever fail at their mission and reliably located or created champions for new technologies. It seemed there were always some outside Microsoft looking to get in early on the next Microsoft technologies. The evangelism team was expert at finding and empowering those people. They could summon the forces of the book, consulting, and training worlds to produce volumes of materials—whole books and training courses on XAML were available even before Vista was broadly deployed. Although WPF had not shipped with any product, it had a strong group of followers who trusted Microsoft and made a major commitment to use WPF for their readers, clients, or customers (as authors, consultants, or enterprise developers). Before Vista shipped, WPF appeared to have initial traction and was a first-class effort, along with .NET on the client. WPF had an internal champion as well. The Zune team used early portions of WPF for software that accompanied their ill-fated, but well-executed, iPod competitor.

Things were less clear when it came to WPF and Vista. WPF code would ship with Vista, but late in the product cycle the shiproom command came down that no one should use WPF in building Vista because of the burden it would place on memory footprint and performance of applications. This caused all sorts of problems when word spread to evangelists (and their champions). People were always looking for signals that Microsoft had changed its mind on some technology. Seeing the risk to WPF by not being performant, the Avalon team (the team was also called Avalon) set out to shrink WPF and XAML into a much smaller runtime—something akin to a more focused and scenario specific product. This was a classic Windows “crisis” project and was added to the product plans sometime in 2005 or early 2006, called Jolt, while the rest of Longhorn was just trying to finish with quality.

Jolt was designed to package up as much of WPF as could fit in a single ActiveX control, also called WPF/E for WPF everywhere, and then later in final form called Silverlight. This would make it super easy to download and use. Streaming videos and small graphical games to be used inside of a browser became the focus. Sound like Adobe Flash? Yes, Jolt was being pitched internally as Microsoft’s answer to Adobe Flash. To compete effectively with Flash, Jolt would also aim to be available across operating systems and browsers—something that made it even less appealing to a Windows strategy, and more difficult to execute. I was of the view that Adobe Flash was on an unsustainable path simply because it was both (!) a fat framework and also an ActiveX control. By this time, ActiveX controls, which a few years earlier were Microsoft’s main browser extensibility strategy, had come to be viewed as entirely negative because they were not supported in other browsers and because they were used as malware by tricking people into running them. The technical press and haters loved to refer to ActiveX as CaptiveX. As an aside, one of my last projects working on C++ was to act overly strategic and push us to adopt the predecessor to ActiveX, known as OLE Controls, and implement those in our C++ library affording great, but useless, synergy with Visual Basic.

For me, this counted as two huge strikes against Jolt. Imagine a strategic project, at this stage in the history of the company, that came about from a crisis moment trying to find any code to ship while also using the one distribution method we had already condemned (for doing exactly the same thing previously.) I did not understand where it was heading.

Somehow, I was supposed to reconcile this collection of issues. When I met with the leaders of the team, they were exhausted though still proud of what they had accomplished. When I say exhausted, I mean physically drained. The struggle they had been through was apparent. Like many who had worked on the three pillars of Vista, the past few years had been enormously difficult. They wanted to salvage something of their hard work. I couldn’t blame them. At the same time, those weren’t the best reasons to ship code to hundreds of thousands of developers and millions of PCs without a long-term strategy for customers. My inclination was to gently shut down this work rather than support it forever knowing there was no roadmap that worked.

The team, however, had done what Windows teams did often—evangelized their work and gained commitments to foreclose any attempt to shut down the effort. With the help of the evangelism team, they had two big customers lined up in addition to the third parties that the evangelism group had secured. In addition to Zune, the reboot of the Windows Phone (which would become Windows Phone 7) would have a portions of the developer strategy based on Jolt—not the phone itself, but it would use Jolt as a way to make it easy for developers to build apps for the phone operating system (prior to this time, apps for the phone were built to basically use ancient Windows APIs that formed the original Windows CE operating system for phones). The Developer division wanted to bring WPF and Indigo into the .NET framework and create one all-encompassing mega-framework for developers but branded as the new version of .NET. The way the .NET framework generally addressed strategic issues was to release a new .NET that contained more stuff under the umbrella of a new version with many new features, even if those new features strategically overlapped with other portions of .NET.

Given all this, the choice was easy for me. As they requested, the Phone team and the Developer division took over responsibility for the Jolt and WPF teams, respectively. It was a no-brainer. Eventually the code shipped with Windows 7 as part of a new .NET framework, which planned anyway. Most everyone on the Windows team, particularly the performance team in COSD and the graphics team in WEX, were quite happy with all of this. The Windows team had always wanted to focus on Win32, even though there was little data to support such a strategy.

While this decision clarified the organization and responsibility, it in no way slowed the ongoing demise of Windows client programming nor did it present a coherent developer strategy for Microsoft. The .NET strategy remained split across WPF and the old .NET client solutions, neither of which had gained or were gaining significant traction—even with so much visible support marshalled by the evangelism team. Win32 had slowed to a crawl, and we saw little by way of new development. It was discouraging.

Again, many reading this today will say they were (or remain) actively developing on one or the other. My intent isn’t to denigrate any specific effort that might be exceedingly important to one developer or customer, but simply to say what we saw happening in total across the ecosystem as evidenced by the data we saw from the in-market telemetry. One of the most difficult challenges with a developer platform is that most developers make one bet on a technology and use it. They do not see histograms or pie charts of usage because they are 100% committed to the technology. They are also vocal and with good reason, their livelihoods depend on the ongoing health of a technology.

With everything to do with developers, APIs, runtimes, and the schism in place, the problem or perhaps solution was this was all “just code,” as we would say. What that means is twofold. First, there was always a way to demonstrate strategy or synergy. In a sense, this was what we’d disparagingly call stupid developer tricks. These were slides, demonstrations, or strategic assertions that showed technical relationships between independently developed, somewhat-overlapping, and often intentionally different technologies. The idea was to prove that the old and new, two competing new, or two only thematically connected technologies were indeed strategically related. My project to support OLE Controls in C++ was such a trick.

Technically these tricks were the ability to switch languages, use two strategies at the same time, or tap into a Win32 feature from within some portion of something called .NET. A classic example of this was during the discussion about where Jolt should reside organizationally. It was pointed out that Jolt had no support for pen computing or subsequently touch (among other things) since there was none in .NET or WPF to begin with. These were both key to Windows 7 planning efforts. Very quickly the team was able to demonstrate that it was entirely possible to directly call Win32 APIs from within Jolt. This was rather tautological, but also would undermine the cross-platform value proposition of Jolt and importantly lacked tools and infrastructure support from Jolt.

Second, this was all “just code,” which meant at any time we could change something and maybe clean up edge cases or enable a new developer trick. Fundamentally, there was no escaping that Win32, .NET, WPF, and even Jolt were designed separately and for different uses with little true architectural synergy. Even to this day, people debate these as developers do—this is simply a Microsoft-only version of what goes on across the internet when people debate languages and runtimes. Enterprise customers expect more synergy, alignment, and execution from a single company. More importantly, developers making a bet on the Microsoft platform expect their choices to be respected, maintained, and enhanced over time. It is essentially impossible to do that when starting from a base as described, and as Microsoft amassed from 2000 to 2007.

As simple as it was to execute moving these teams, in many ways it represented a failure on my part, or, more correctly, a capitulation. I often ask myself if it would have been better to wind down the efforts rather than allow the problem to move to another part of the company. I was anxious to focus and move on, but it is not clear that was the best strategy when given an opportunity to bring change and clarity when developers so clearly were asking for it, even if it meant short-term pain and difficulty. It would have been brutal.

It grew increasingly clear that there were no developers (in any significant number, not literally zero) building new applications for Windows, not in .NET or WPF or Win32. The flagship applications for Windows, like Word, Excel, and others from Microsoft along with Adobe Photoshop or Autodesk AutoCAD, were all massive businesses, but legacy code. The thousands of IT applications connecting to corporate systems, written in primarily Visual Basic, all continued daily use but were being phased out in favor of browser-based solutions for easier deployment, management, portability, and maintenance. They were not using new Windows features, if any existed. The most active area for Windows development was gaming, and those APIs were yet another layer in Windows, called DirectX, which was part of WEX and probably the most robust and interesting part of Win32. Ironically, WPF was also an abstraction on top of those APIs. ISVs weren’t using anything new from Microsoft, so it wasn’t as though we had a winning strategy to pick from.

Further evidence of the demise of Win32 arose as early as 1996 from Oracle’s Larry Ellison, who put forth the idea of a new type of browser-only computer, the Network Computer. (Sound like a Chromebook?) At the time, Marc Andreessen famously said that Windows would be become, simply, a “poorly debugged set of device drivers,” meaning that the only thing people would care about would be running the browser. Years later, Andreessen would point out the original was based on something Bob Metcalf had said. Eight years later we had reached the point where the only new interesting mainstream Windows programs were browsers, and at this time only Firefox was iterating on Windows and the only interesting thing about Firefox was what it did with HTML, not Windows. The device drivers still had problems in Vista. In fact, that was the root of the Vista Capable lawsuit!

Moving WPF and Jolt to their respective teams, while an admission of defeat on my part, could best be characterized as a pocket veto. These were not the future of Windows development, but I wasn’t sure what would be. We in Windows were doubling down on the browser for sure, but not as leaders, rather as followers. We had our hands full trying to debug the device drivers.

XAML development continues today, though in a much different form. While it does not make a showing in the widely respected Stack Overflow developer survey of over 80,000 developers in 181 countries, it maintains a spot in the Microsoft toolkit. XAML will come to play an important role in the development of the next Windows release as well.

With the team on firm(er) ground and now moving forward we finally started to feel as though we had gained some control. By September 2007 we were in beta test with the first service pack for Vista, which OEMs and end-users anxiously awaited.

The team was in full execution mode now and we had milestones to plow through. While I felt we were heading in the right direction and cleared the decks of obvious roadblocks, there was a looming problem again from Cupertino. What was once a side bet for Microsoft would prove to be the most transformative invention of the 21st century…from Apple.

On to 092. Platform Disruption…While Building Windows 7 [Ch. XIII]

Advertising is much more difficult than just about everyone believes to be the case. In fact, one of the most challenging tasks for any executive at any company is to step back and not get involved in advertising. It is so easy to have opinions on ads and really randomize the process. It is easy to see why. Most of us buy stuff and therefore consume advertising. So it logically follows, we all have informed opinions, which is not really the case at all. Just like product people hate everyone having opinions on features, marketing people are loathe to deal with a cacophony of anecdotes from those on the sidelines. Nothing would test this more for all of Microsoft than Apple’s latest campaign that started in 2006. I’d already gone through enough of watching advertising people get conflicting and unreconcilable feedback to know not to stick my nose in the process.

Back to 089. Rebooting the PC Ecosystem

Hardcore Software by Steven Sinofsky is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

“I’m a Mac.”

“And, I’m a PC.”

The “Get a Mac” commercials, starting in 2006, changed the competitive narrative overnight and were a painful gut punch welcoming me to Windows.

They were edgy, brutal in execution, and they skewered Windows with facts. They were well done. (Though it always bothered me that PC Guy had a vague similarity to the late and much-loved Paul Allen, Microsoft’s cofounder long focused on science and philanthropy.) Things that drove Windows fans crazy like “no viruses” on Mac were not technically true, but true in practice because, mostly, why bother writing viruses for Macs with their 6 percent share? That’s what we told ourselves. In short, these commercials were devastating.

They probably bumped right up against a line, but they effectively tapped into much of the angst and many of the realities of PC ownership and management. Our COO and others were quite frustrated by them and believed the commercials not only to be untrue, but perhaps in violation of FTC rules.

They were great advertising. Great advertising is something Apple seemed to routinely accomplish, while Microsoft found it to be an elusive skill.

For its first twenty years, Microsoft resisted broad advertising. The company routinely placed print ads in trade publications and enthusiast magazines, with an occasional national newspaper buy for launches. These ads were about software features, power, and capabilities. Rarely, if ever, did Microsoft appeal to emotions of buyers. When Microsoft appeared in the national press, it was Bill Gates as the successful technology “whiz kid” along with commentary on the growing influence and scale of the company.

With that growing influence in the early 1990s and a business need to move beyond BillG, a huge decision was made to go big in advertising. Microsoft retained Wieden+Kennedy, the Portland-based advertising agency responsible for the “Just Do It” campaign from Nike, among many era-defining successes. After much consternation about spending heavily on television advertising, Microsoft launched the “Where do you want to go today?” campaign in November 1994.

Almost immediately we learned the challenges of advertising. The subsidiaries were not enamored with the tagline. The head of Microsoft Brazil famously pushed back on the tagline, saying the translation amounted to saying, “do you want to go to the beach today?” because the answer to the question “Where do you want to go?” in Brazil was always “the beach.” The feedback poured in before we even started. It was as much about the execution as the newness of television advertising. Everyone had an opinion.

I remember vividly the many pitches and discussions about the ads. I can see the result of those meetings today as I rewatch the flagship commercial. Microsoft kept pushing “more product” and “show features” and the team from W+K would push back, describing emotions and themes. The client always wins, and it was a valuable lesson for me. Another a valuable lesson, Mike Maples (MikeMap) who had seen it all at IBM pointed out just before the formal go ahead saying something like “just remember, once you start advertising spend you can never stop. . .with the amount of money we are proposing we could hire people in every computer store to sell Windows 95 with much more emotion and information. . .” These were such wise words, as was routine for Mike. He was right. You can never stop. TV advertising spend for a big company, once started, becomes part of the baseline P&L like a tax on earnings.

The commercials were meant to show people around the world using PCs, but instead came across almost cold, dark, and ominous as many were starting to perceive Microsoft.  That was version 1.0. Over the next few years, the campaign would get updates with more colors, more whimsy, and often more screenshots and pixels.

What followed was the most successful campaign the company would arguably execute, the Windows 95 launch. For the next decade, Microsoft would continue to spend heavily, hundreds of millions of dollars per year, though little of that would resonate. Coincident with the lukewarm reception to advertising would be Microsoft’s challenges in branding, naming, and in general balancing speeds and feeds with an emotional appeal to consumers. Meanwhile, our enterprise muscle continued to grow as we became leaders in articulating strategy, architecture, and business value.

In contrast, Apple had proven masterful at consumer advertising. From the original 1984 Superbowl ad through the innovative “What’s on your PowerBook?” (1992) to “Think Different” (1997-2000) and many of the most talked about advertisements of the day such as “C:\ONGRTLNS.W95” in 1995 and “Welcome, IBM. Seriously” in 1981, Apple had shown a unique ability to get the perfect message across. The only problem was that their advertising didn’t appear to work, at least as measured by sales and/or market share. The advertising world didn’t notice that small detail. We did.

Starting in 2006 (Vista released in January 2007), Apple’s latest campaign, “Get a Mac” created an instant emotional bond with everyone struggling with their Windows PC at home or work, while also playing on all the stereotypes that existed in the Windows v. Mac battle—the nerdy businessman PC slave versus the too cool hipster Mac user.

The campaign started just as I joined Windows. I began tracking the commercials in a spreadsheet, recording the content and believability of each while highlighting those I thought particularly painful in one dimension or another. (A Wikipedia article would emerge with a complete list, emphasizing the importance of the commercials.) I found myself making the case that the commercials reflected the state of Windows as experienced in the real world. It wasn’t really all that important if Mac was better, because what resonated was the fragility of the PC. There was a defensiveness across the company, a feeling of how the “5% share Mac” could be making these claims. I managed a bit of a row with the COO who wanted to go to the FTC and complain that Apple was not telling the truth.

Windows Vista dropped the ball. Apple was there to pick it up. Not only with TV commercials and ads, but with a resurging and innovative product line, one riding on the coattails of Wintel. The irony that the commercials held up even with the transition to Intel and a theoretically level playing field only emphasized that the issue was software first and foremost, not simply a sleek aluminum case.

While the MacBook Air was a painful reminder of the consumer offerings of Windows PCs, the commercials were simply brutal when it came to Vista. There were over 50 commercials that ran from 2006-2009, starting with Apple’s transition to Intel and then right up until the release of Windows 7 when a new commercial ran on the eve of the Windows 7 launch. Perhaps the legacy of the commercials was the idea that PCs have viruses and malware and Macs do not. No “talking points” about market share or that malware targets the greatest number of potential victims or simply that the claim was false would matter. There’s no holding back, this was a brutal takedown, and it was effective. It was more effective in reputation bashing, however, than in shifting unit share.

One of the most memorable ones for me was “Security” which highlighted the Windows Vista feature designed to prevent viruses and malware from sneaking on your PC, called User Account Control or UAC which had become a symbol of the annoyance of Vista—so much so our sales leaders wanted us to issue a product change to remove it. There’s some irony in that this very feature is not only implemented across Apple’s software line today, but far is more granular and invasive. That should sink in. Competitively we all seem to become what we first mock.

Something SteveB always said when faced with sales blaming the product group and the product group blaming sales for something that wasn’t working, was “we need to build what we can sell, and sell what we build.” Windows Vista was a time where we had the product and simply needed to sell what we had built, no matter what.

The marketing team (an organization peer to product development at this time) was under a great deal of pressure to turn around the perceptions of Vista and do something about the Apple commercials. It was a tall order. The OEMs were in a panic. It would require a certain level of bravery to continue to promote Vista, perhaps not unlike shipping Vista in the first place. The fact that the world was in the midst of what would become known as the Global Financial Crisis with PC sales taking a dive did not help.

Through a series of exercises to better come up with a point of attack the team came up with the idea that maybe too many people were basing their views of Windows Vista on hearsay and not actual experience. We launched a new campaign “Mojave Experiment” that took a cinéma vérité approach to showing usability studies and focus groups for yet to be released and experimental version of Windows Mojave. After unsolicited expressions of how bad Vista was, the subjects were given a tour of Windows Mojave. The videos were not scripted, and the people were all real, as were the reactions. Throughout the case study, and the associated online campaign, the subjects loved what they were being shown. Then at the end of the videos, the subjects were told the surprise—this was Windows Vista. To those old enough to know, it had elements of the old “Folger’s Coffee” or “I Can’t Believe It’s Not Butter” commercials, classic taste-test advertising.

The industry wasn’t impressed. In fact, many took to blog posts to buzzsaw the structure of the tests and way subjects were questioned and shown the product. One post in a Canadian publication ran with the headline “Microsoft thinks you’re stupid” in describing the campaign. This was right on the heels of the Office campaign where we called our customers dinosaurs. We had not yet figured out consumer advertising.

We still had to sell the Vista we had built. We needed an approach that at the very least was credible and not embarrassing, but importantly at least hit on the well-known points about Apple Macintosh that everyone knew to be true. Macs were expensive and as a customer your choices were limited.

Apple’s transition to Intel was fascinating and extraordinarily well executed, releasing PCs that were widely praised, featuring state-of-the-art components, an Intel processor unique to Apple at launch, and superior in design and construction to any Windows PC. The new premium priced Intel Macs featured huge and solid trackpads, reliable standby and resume, and super-fast start-up. All things most every Windows PC struggled to get right. We consistently found ourselves debating the futility of the Apple strategy offering expensive hardware. The OEMs weren’t the only ones who consistently believed cheaper was better, but that was also baked into how Microsoft viewed the PC market. Apple had no interest in a race to profitless price floors and low margins, happily ceding that part of the market to Windows while selling premium PCs at relatively premium prices. In fact, their answer to the continued lowering of PC prices was to release a pricey premium PC.

The original MacBook Air retailed for what seemed like an astonishing $1,799. That was for the lowest specification which included a 13” screen and a meager 2GB main memory, an 80GB mechanical hard drive, and a single USB port and obscure video output port, without an DVD drive or network port. For an additional $1000, one could upgrade to a fancy new solid state disk drive which was still unheard of on mainstream Windows PCs.

As it would turn out the MacBook Air was right in the middle of the PC market, and that’s just how PC makers liked it, stuck between the volume PC and premium PC, neither here nor there. An example of most popular laptop configuration was the Dell Inspiron 1325. The 1325 was a widely praised “entry level” laptop with an array of features, specs, and prices. In fact, on paper many PC publications asked why anyone would buy an overpriced Macintosh. The Dell 1325 ranged in prices from $599 to about $999 depending on how it was configured. The configuration comparable to a MacBook Air was about $699 and still had 50% more memory and three times the disk space. As far as flexibility and ports, the 1325 featured not just a single USB port but two, a VGA video connector, audio jacks, Firewire (for an iPod!), an 8 in 1 media card reader, and even something called an ExpressCard for high-speed peripherals. Still, it was a beast, while the same width and length it was twice as thick and clearly more dense weighing almost 5lbs at the base configuration compared to the 3lb MacBook Air. As far as battery life, if you wanted to be comparable to the Air then you added a protruding battery that added about a pound in weight and made it so the laptop wouldn’t fit in a bag. Purists would compare to the MacBook (not the Air), as we did in our competitive efforts, but the excitement was around the Air. The regular 13” MacBook weighed about 4.5lbs and cost $1299, which would make it a more favorable comparison. It was clear to me that the Air was the future consumer PC as most PC users would benefit from lighter weight, fewer ports, and a simpler design. As much as I believed this, it would take years before the PC industry broadly recognized that “thin and light” was not a premium product. The MacBook Air would soon end up priced at a $999 entry price, which is when it began to cause real trouble for Windows PCs.

The higher-end MacBook Air competitor from the PC world was the premium M-series from Dell. Incidentally, I’m using Dell as an example, HP and Lenovo would be similar in most every respect. The Dell XPS M1330, the forerunner to today’s wonderful Dell XPS 13, was a sleeker 4lbs also featuring a wedge shape. With the larger and heavier battery there was a good 5 hours of runtime. Both Dells featured plastic cases with choices of colors. It too had models cheaper than the MacBook or MacBook Air but could be priced significantly more by adding more memory, disk storage, better graphics, or a faster CPU.

A key factor in the ability for the Mac to become mainstream, however, was the rise in the use of the web browser for most consumer scenarios. A well-known XKCD cartoon featured two stick figures, one claiming to be a PC and another claiming to be a Mac, with the joint text pointing out “and since you do everything through a browser now, were pretty indistinguishable.” Apple benefitted enormously from this shift, or disruption, especially as Microsoft continued to invest heavily in Office for the Mac. The decline in new and exciting Windows-based software described in the previous section proved enormously beneficial to Apple when it came to head-to-head evaluation of Mac versus Windows. Simply running Office and having a great browser, combined with the well-integrated Apple software for photos, music, videos, and mail proved formidable, and somewhat enduring with the rise in non-PC devices.

We were obsessed with the pricing differences. We often referred to these higher prices as an “Apple Tax” and even commissioned a third party to study the additional out of pocket expenses for a typical family when buying Macs versus Windows PCs. A whitepaper was distributed with detailed comparison specifications showing the better value PCs offered. In April 2008 we released a fake tax form itemizing (groan) the high cost of Apple hardware.

From our perspective or perhaps rationalization this was all good. Consumers had choice, options, and flexibility. They could get the PC they needed for their work and pay appropriately, or not. This thesis was reinforced by the sales of both PCs and Macs no matter what anyone was saying in blogs. The PC press loved this flexibility. Retailers and OEMs relied on the variety of choices to maximize margin. Retailers in particular struggled with Apple products because they lacked key ways to attach additional margin, such as upsell or service contracts, not to mention Apple’s lack of responsiveness to paying hefty slotting and co-advertising fees.

Choosing a PC while a complicated endeavor was also the heart and soul of the PC ecosystem. Once Apple switched to Intel, there was a broad view that the primary difference between Mac and Windows now boiled down to the lack of choice and high price and lack of a compatible software and peripheral ecosystem that characterized Macintosh.

To make this point, Microsoft launched a new campaign the “Laptop Hunter” that ran in 2009. In these ads, typical people are confronted outside big box retailers trying to decide what computer to buy. A PC or a Mac? In one ad, “Lauren” even confesses she is not cool enough for a Mac while noticing just how expensive they are (NB, Lauren is almost the perfect representation of a Mac owner.) She heads over to a showroom with a vast number of choices and whittles her way down to a sub $1000 PC with everything she needs. Another success.

Not to belabor this emerging theme, but no one believed these ads either. Only this time, the critics and skeptics were livid as it appeared Lauren was an actress and that called into question the whole campaign. In addition, Apple blogs went frame-by-frame in the ad to “prove” various aspects of the shoot were staged or simply not credible. The tech blogs pointed out the inconsistencies or staged aspects of Lauren’s requirements as being designed to carefully navigate the Apple product line.

Laptop Hunter offers some insight into how addictive (or necessary) television advertising became and the scale at which Microsoft was engaging. The television campaign debuted during the NCAA basketball tournament in the US, prime time. It was supported by top quality network prime shows (Grey's Anatomy, CSI, The Office, Lost, American Idol), late night staple programming (Leno, Letterman, Kimmel, Conan, and Saturday Night Live) and major sports events and playoff series (NCAA Basketball, the NBA, MLB, and the NHL). Cable networks included Comedy Central, Discovery, MTV, VH1, History and ESPN. The online campaign included home page execution on NYTimes.com, as well as “homepage take-overs” (the thing to do back then) on WSJ, Engadget, and CNN. We also supported this with an online media buy targeted at reaching people who were considering a non-Windows laptop (a Mac). We linked from those banner ads to a dedicated Microsoft web site designed to configure a new PC and direct to online resellers, closing the loop to purchase. The level of spending and effort was as massive as the upside.

I could defend the advertising but at this point I am not sure it is worth the words. Besides, Apple responded to the ad with a brutal reiteration of viruses and crashes in PCs and that lots of bad choice is really no choice at all. It is rare to see two large companies go head-to-head in advertising and you can see how Microsoft took the high road relative to Apple, deliberately so. The ads worked for Apple, but almost imperceptibly so in the broader market. Apple gained about one point of market share, which represented over 35% growth year over year for each of the two years of the campaign—that is huge. The PC market continued to grow, though at just over 10%. Still that was enough of a gain to ameliorate the share gains from Apple, which were mostly limited to the US and western Europe.

As much as the blowback from this campaign hurt, we were at least hitting a nerve with Apple fans and getting closer to a message that resonated with the PC industry: compatibility, choice, and value.

For our ad agency, essence of the PC versus Mac debate boiled down not to specs and prices but to a difference in the perceived customers. The Mac customers (also the agency itself) seemed to be cut from one mold of young, hip, artistic whereas the PC was literally everyone else. It seemed weird to us, and our advertising agency, that Windows computers were not given credit for the wide array of people and uses they supported, if even stereotypically. We were proud of all the ways PCs were used.

To demonstrate this pride, Bill Veghte (BillV) the senior vice president of Windows marketing (also reporting to Kevin Johnson) led the creation of a new “I’m a PC” campaign that started in the fall of 2008 and ran through the launch of Windows 7. Rather than run from the Apple’s “I’m a Mac” we embraced it. The main spot featured fast cuts of people from all walks of life, including members of the Microsoft community as well as some pretty famous people, talking about their work, creations, and what they do with PCs. The ads featured a Microsoft employee, Sean Siler a networking specialist from Microsoft Federal, who looked unsurprisingly like the stereotype PC users portrayed by Apple. These ads were us.

The advertising world viewed success through the creative lens so dominated by Apple. The ads were well-received and for the first time we landed spots (costing hundreds of millions of dollars) that we could both be proud while emphasizing our strengths.

The memorable legacy of the campaign would be the brightly colored “I’m a PC” stickers that nearly everyone at the company dutifully attached to their laptops. Meeting rooms filled with open laptops of all brands, colors, sizes all displayed the sticker. We made sure all of our demo machines featured the stickers as well. In the summer 2009 global sales meeting just before Windows 7 would launch, BillV led the sales force in a passionate rally around “I’m a PC” and the field loved it. He was in his element, and they were pumped. The Windows 7 focused versions of this campaign featured individuals talking about their work saying “I’m a PC and Windows 7 was my idea” building on the theme of how Windows 7 better addressed customer needs (more on that in the next chapter.)

By the summer of 2009, the back and forth with Apple seemed to run its course as we were close to the launch of Windows 7 (more on that in the next chapter.) The New York Times ran a 3000-word, front of the Sunday business section story titled “Hey, , PC, Who Taught You to Fight Back?” covering what was portrayed as an “ad war, one destined to go down in history with the cola wars of the 1980s and ’90s and the Hertz-Avis feud of the 1960s.” There was even a chart detailing the escalating advertising spend of the two companies. The story even noted that the ads caught the attention of Apple who pulled their ads from the market only to return with new “Get a Mac” ads criticizing Microsoft’s ads. In the world of advertising, that counts as a huge victory. On the store shelves, the campaign finally seemed to at least slow the share loss to Apple worldwide and definitely pushed it back in the US.

Nothing hit home more a few years later than the photo of the White House situation room in May 2011 during the raid on Osama Bin Laden. That photo was captioned by the internet to illustrate the point of just who is a Mac and who is a PC. The meme featured some barefooted hipsters in a coffee shop captioned “I’m a Mac” and then the situation room featured secured PCs captioned “I’m a PC” with the heading “Any Questions?” We loved it. That seriousness was what we were all about.

Of course, the real battle with Apple was now about software. Windows 7 needed to execute. We needed to build out our services offerings for mail, calendar, storage, and more where Apple was still flailing even more than we were.

While I was entirely focused on Windows 7 and moving forward, the ghosts of Vista and Longhorn would appear. Promises were made that ultimately could not be kept and we had to work through those.

On to 091. Cleaning Up Longhorn and Vista

Postscript. The “Get a Mac” ad that hit me the hardest for non-product reasons was the “Yoga” spot, which was funny to me because when I moved to Windows in March 2006 I switched to practicing yoga after a decade of Pilates. In the spot, PC guy switches from yoga to Pilates.

The word ecosystem is often used when describing Windows and the universe of companies that come together to deliver Windows PCs and software. Providing a platform is a much trickier business than most might believe. Bringing together a large number of partners, along with their competitors, who might share one large goal but differ significantly in the tactics to use to achieve it is fraught with conflict. The Windows ecosystem had been dealt a series of painful blows over the years resulting in a loss of trust and collective capability. Where partnerships were required, the ecosystem had become a collection of…adversaries…or direct competitors…or conflicting distribution channels.

Back to 088. Planning the Most Important Windows Ever

In the summer of 2007 six-months after Windows Vista availability as we rolled out the Windows 7 product vision to the entire team, the press coverage for Vista began to get brutal. I say the coverage, but this is really about the customer reaction to the product and the press simply reflected that.

Then the OEMs, large public companies with shareholders and quarterly earnings, started to do something almost unimaginable. They began to speak out about the problems with Vista. In a widely covered interview in the Financial Times discussing quarterly results, Acer president Gianfranco Lanci lashed out at Vista, saying the “whole industry is disappointed with Windows Vista" and stated that Vista had stability problems and he doubted that Microsoft would remedy the issues within the next six months. He went on to suggest that customers really wanted Windows XP knowing that in just a few months XP would be discontinued. Much of what he said in public, the OEMs had been telling us in private.

I had been in my role for more than a year and still did not have much to say about what was coming next. How could I? We just didn’t know and all my experience with customers told me that claiming I didn’t know would not be acceptable or even credible. Still, pressure was mounting to show progress.

The Windows (or PC) ecosystem is made up Intel, PC makers (the OEMs), and creators hardware components and peripherals known as Independent Hardware Vendors (IHVs). OEMs accounted for the vast majority of Microsoft’s extremely lucrative Windows business. IHVs were the key ingredients the OEMs counted on for innovation. Intel, provider of CPUs and an increasing portion of the main componentry, was the most central in the hardware ecosystem as half of the legendary Wintel partnership.

Suffice it to say ecosystem was in an unhappy and untrusting state after years of product delays along with a series of feature, product, and pricing miscues. The delays had been extremely painful to OEMs and Intel, who counted on a new release of Windows to show off new PCs and drive growth in PC sales. IHVs had critical work to do enabling new PCs. They had to build software drivers that enabled new hardware that was compatible with the new features of Windows (such as 64-bit Windows or new security features). They also faced the increasing complexity of maintaining drivers for old versions of Windows as the time between releases increased and customers demanded new hardware support on both old and new platforms. The rise of Linux was spreading the Windows ecosystem further as demand for Linux support increased. Fractures were everywhere.

Intel contributed an increasingly larger portion of the PC, much as Microsoft continued to add features to Windows. With Centrino® for example, Intel added WiFi support directly to the components they provided to PC makers. This greatly expanded and standardized the use of WiFi in laptops. Intel was in the process of broadening their support for graphics as well, continuing to improve the integrated graphics chips they provided to PC makers (this was a particularly sore spot with Vista.) Intel was able to better control pricing of their components, encourage specific PC designs, and shift PC makers to various CPU choices using a combination of pricing actions and co-marketing arrangements. Originally these “Intel Inside” advertising efforts were a huge part of how PC makers selected and marketed different models and lines of computers and was enormously profitable for Intel.

Microsoft’s relationship with Intel had not been particularly happy, going way back to the 1990s when Intel began to embrace Java and other cross-platform technologies. Recently, however, the rise of Linux was viewed by Intel as a large opportunity, while Microsoft perceived it as a competitive threat. What Microsoft used to perceive as a moat, device driver support for Windows, was rapidly fading due to active efforts by Intel to support Linux to the same degree. The specter of desktop Linux seemed to be held back by just a few device drivers should Intel support them, or so we worried.

There was also a complex relationship between enterprise customers and Microsoft when it came to Windows. From a product and feature perspective, Windows was making increasing bets on the kind of features businesses cared about, such as endpoint control, reliability, and management. Microsoft did not always want to give away those features for free to retail customers where they might not apply. The fact that these features existed caused OEMs to consider ways to add them to the base Windows they sold, feeling that the base Windows was under-powered.  In other words, features Microsoft added to premium editions of Windows simply served as starting points for features OEMs might look to provide with their own software on top of the basic editions of Windows, increasing OEM margin.

The response to the increasing gap between consumers and enterprise was another reason for OEMs to embrace, or at least appear to embrace Linux. At the very least, OEMs began to ponder the idea of selling PCs with free Linux and letting customers pick and choose an operating system on their own as a backdoor way to avoid the Windows license for every single PC. Microsoft did not want OEMs selling PCs without Windows when Windows was destined to be on the PC anyway, as that would only encourage piracy. This “Linux threat” was not as empty as the state of technology would have indicated as OEMs were actively starting to offer Linux on the desktop, especially if they were offering it on Server already. Some markets, such as China, which already had enormously high piracy pursued this path aggressively.

From the least expensive PCs for the budget conscious to the fanciest PCs for gamers, the Windows product and business depended on these partners, and vice versa. Any business book would tell you just how big a deal the Windows ecosystem was, as would anyone who followed the antitrust actions against Microsoft.

We needed to reboot the relationship with and across the Ecosystem. Doing so would be an important part of the planning process and run in parallel with planning the Windows 7 product.

The perspectives, insights, and data from hardware and OEM partners were only part of the input to the planning process. Another part was the usage data from tens of millions of Windows users. What features, peripherals, and third-party programs were used, how often, and by which types of customers. Vista had done excellent work to incorporate the performance and reliability measures of the system, especially for glitches like crashes and hangs, but usage data was inconsistent across the system. As we would discover, this data was not readily available or reliably collected across the whole product. We knew we would need as much as we could for Windows 7 and certainly down the road, so we began in earnest to implement more telemetry across the product during the planning for Windows 7. This data would form a foundation for many discussions with OEMs.

The feverish pace and tons of work over the six months that followed the re-org tackled an expansive set of potential areas and distilled them down to a product plan, the Windows 7 Vision. That six months would bump up against not one but two OEM selling seasons during which the OEMs would not get the news about “fixing Vista” they were hoping for. The planning memo came out in December 2006 even before Windows Vista availability. That timing, however, would be too late to impact the February selling season. The July 2007 Vision was too late to become requirements for PCs for back to school in August or September of that year. Even though PCs would not ship with Windows 7 for quite some time, every selling season we missed would make it frustratingly difficult for those PCs to be upgraded to Windows 7. This was a problem for every part of the ecosystem including Microsoft.

The relationships were difficult, but more importantly the quality of work across the ecosystem was in decline. Communication was adversarial at best.

Mike Angiulo (MikeAng) joined from Office and began an incredible effort at rebuilding the relationships with OEMs. Mike was not only a stellar technologist, but a strategic relationship manager, having grown up as a salesman within his future family’s business. He was also a trained mechanical engineer who rebuilt and raced cars, a natural-born poker champion, and an IR-pilot who built his own plane.

He recruited Roanne Sones (RSones) to lead the rehabilitation of the relationship with OEM customers. She was a college hire from Waterloo’s prestigious systems engineering program who had joined Office five years earlier. Having worked on projects from creating the Office layperson’s specification, to synthesizing customer needs by segment, to analyzing usage data across the product, Roanne brought a breadth of tools and techniques from Office to the OEM problem space. Also joining the team was Bernardo Caldas (BCaldas) who would bring deep insights combining usage data, financial modeling, and business model thinking to the team.

Roanne quickly learned that a big part of the challenge would be the varying planning horizons the OEMs required. “Required” because they told us that they needed “final Windows 7 plans” in short order if they were to make the deadlines for a selling season. Whether they needed the information or not, the reality was they were so used to not getting what they wanted they simply made any request urgent and expansive.

Each OEM was slightly different in how it approached the relationship, but all shared two key attributes. First, they viewed Windows as an adversary. Second, they did not take anything we told them to be the ground truth. They were constantly working the Microsoft organization and their connections, which were deep, across the separately organized OEM sales team, support organizations, and more, to triangulate anything we said to arrive at their own version of truth or reality. They also had no problems escalating to the head of global sales or to SteveB directly, several of whom had known Steve for decades.

After a decade of missed milestones, dropped features, antitrust concerns, and contentious business relationships, the connection between Microsoft and OEMs was painfully dysfunctional. Considering Windows accounted for so much of Microsoft’s profit, this was a disaster of a situation. It had been going on so long that most inside Microsoft seemed to shrug it off as just a part of the business or “OEMs have always been like this.” It was no surprise to me that the relationships devolved.

We found little to love in most new PCs while at the same time Microsoft was rooted in a long history of believing, as BillG would say though this is misinterpreted I think, that “ten years out, in terms of actual hardware costs you can almost think of hardware as being free.” He meant that to imply that relative to advancing software capabilities, the hardware resources required would not be a key factor in innovation. I’m not sure OEMs heard it that way.

What was really going on other than Microsoft’s lack of reliability as a partner, not to dismiss that perfectly legitimate fact? It goes back to the history of PC manufacturing and its evolution to the current situation. When Dell and then Compaq first manufactured PCs, they were rooted in engineering and design of PCs, and each built out manufacturing and distribution channels—PCs made in the US (Texas) and shipped around the world.

As the industry matured, more and more manufacturing moved to plants in Taiwan and China. Traditionally, components such as hard drives, accessory boards, and motherboards were shipped to the United States from locations around Asia for final assembly, including the addition of the Intel CPU manufactured in the US. Eventually, as more and more were made in Asia, it became increasingly efficient to aggregate components there for assembly as a complete PC, which could then be shipped around the world. This transition is one Tim Cook, now the CEO of Apple, famously took Apple through even though Steve Jobs resisted it.

Over time, these assembly companies aimed to deliver even more value and a greater share of the PC experience, as they described. They even began to create speculative PCs and sell them to the OEMs, such as Dell and HP for incorporation into their product lines after a bit of customization such as component choice and industrial design. The preference for laptops over desktops made it even more critical to engineer a complete package, which these new original design manufacturer partners became experts at doing. As the name implied, an ODM would both design and manufacture PCs (and many other silicon-based electronics). The ODMs developed such complete operations that some even created consumer-facing brands to sell PCs as a first party, in search of margins and revenue.

In a sense, the headquarters of an OEM was the business operation and the ODM the product and manufacturing arm, managed on metrics of cost, time to delivery, and quality. In this view, Windows itself became just another part of the supply chain, albeit the second most expensive part, usually far behind Intel. That’s why PCs tended to converge on similar designs. Some argued that the ODM process drove much of that with a small set of vendors looking to keep costs low and sourcing from a small set of suppliers to meet similar needs from US sales and marketing arms. Even though there were a dozen global PC makers, the increasing level of componentization and the ODM model caused a convergence, first with desktops and now we were seeing the same happen with laptops.

Effectively, ODMs were driving a level of commoditization. On the one hand, this was great for better device driver support and a more consistent Windows experience, as investments the ecosystem made were leveraged across independent PC makers. On the other hand, this led to margin compression for PC makers which put even more pressure on ODMs and Microsoft. Further, given all PC makers faced similar constraints, PCs tended to converge to an average product, rather than an innovative one. There were a few outliers such as Sony who continued to drive design wins, but with ever-decreasing volume (Sony sold off their PC business in 2014.)

Roanne and Bernardo were treating the ODMs with the same level of attention as OEMs, something that had not been done before. The PC business was extremely healthy in terms of unit volumes, but the OEMs were all struggling to maintain profit margins and were looking to ODMs for leverage. There was a great deal of envy of Apple and its sleek new MacBook laptops made from machined aluminum, but, more importantly, the margins Apple earned from those PCs were impressive. The OEMs were pressuring the ODMs to deliver the same build quality with room for ample margins and a lower price than Apple, which was believed to be charging premium prices at much lower volume. In manufacturing, volume is everything so it is reasonable to assume that much higher volume could make lower prices possible, but only if the volume was for a small number of different models which the OEMs were not committed to. The ODMs thought this was impossible and continued to push their ability to deliver higher-margin and higher-priced premium laptops.

Early in my time at Windows, I went to Asia to visit with some of the ODMs to see firsthand their perspective on the PC and the relationships across the ecosystem. Visiting with the ODMs brought these tension front and center and proved informative.

The ODMs that served multiple OEMs were quite stringent about secrecy and leaks across their own companies. Each OEM had distinct and secured facilities, and the ODM management structure was such that information was not shared across facilities. Visiting a single ODM meant seeing buildings for any one of the major global manufacturers all identical on the outside, but each entrance was guarded as though the building was owned and operated by the specific OEM. Access was closely guarded and employees of the ODM never crossed into different facilities. I recall once having an Apple building pointed to in a manner that reminded me of driving by a building on the Washington, DC Beltway when everyone knows it is a spy building, but just remains quiet and doesn’t say a word. An ODM even acknowledging they served Apple would jeopardize their business even though it was an open secret.

The ODMs themselves were struggling with margins, even though they benefitted from a highly favorable labor cost structure in Asia. Several were responsible for manufacturing Apple devices, and while they would certainly never, ever divulge any details about that process, we knew that much of the math they would show us, bleak as it was, was informed by what they were doing for Apple and other OEMs.

One of the larger ODMs that I also understood made Macintosh laptops—and was run by a founder and CEO who had grown the company from the 1980s—proved to be an adventure to visit.

I arrived for my tour and put on a bunny suit and booties, deposited any electronics and possessions into a safe, and entered through a metal detector. There were cameras everywhere. There were acres of stations on the floor of a massive assembly line, each one responsible for a step in the manufacturing process. Pallets of hard drives, motherboards, screens, and cases arrived in one end, and progressed through the enormous assembly line stopping at each station. One added the hard drive, while one was attached the screen. Another verified that everything powered up. The final assembly step was where they attached the Genuine Windows hologram that proved that the Windows software was not pirated and was purchased legally. The ODMs always made a point of showing me that step knowing their role in anti-piracy was something we were always on the lookout for. Then the machines were powered up and burned in for a few hours to make sure everything worked.

After seeing the line, we ventured to the top-floor of the building to the CEO’s private office, which was the entire floor. On this floor was a beautiful private art collection of ancient Chinese calligraphy and watercolors. After a ritual of tea and tour of the gallery, the discussions of business began.

Instead of hearing about all the requirements from Windows, the CEO raised many issues about our mutual customers. We discussed the constant squeeze for margin, the pressure to compete with Apple without paying Apple prices, and what I found most interesting: the desire for flexibility, which precluded everything else. The OEMs, it turned out, were born out of an era during which desktop PCs were made from a number of key peripherals, such as a hard drive, graphics card, memory, and other input/output cards. Each of these could be chosen and configured at time of purchase. This allowed for two crucial elements of the business. First, the customer had a build-to-order mindset, which was appealing and a huge part of the success of Dell. Second, the OEM could bid commodity suppliers of these components against each other and routinely swap out cheaper parts while maintaining the price for customers. This just-in-time manufacturing and flexible supply chain were all the rage in business schools.

The problem was that this method did not really work for laptops and especially did not work for competing with Apple. Apple designed every aspect of a laptop and chose all the components and points of consumer choice up front. This gave Apple laptops the huge advantage of being smaller and lighter while also not having to account for a variety of components placing different requirements on software, cooling, or battery life. This integrated engineering was almost the exact opposite of how the Windows PC makers were designing laptops.

During one visit in 2008, after the MacBook Air had recently been released, it was all the ODMs could talk about. The only PCs that came close were made in Japan for the Japan market or achieved little critical mass in the US except among tech elites, such as the Sony VAIO PCG-505 or the Fujitsu LifeBook that I used. The Air’s relatively low(-ish) price point would eventually lead to an Intel marketing initiative called Ultrabook™ but as future sections will describe, it would be years before the PC ecosystem could respond to the Air.

From an ODM perspective, the requirements for the supply chain were driven by non-engineers or marketing from the OEMs back at HQ and seemed disconnected from the realities of manufacturing. They felt they had the capability to build much sleeker PCs than they were being asked. Always implied but never stated was the fact that some of them built devices for Apple.

Meeting after meeting, I heard stories of ODMs who knew how to build leading and competitive laptops but the US OEMs, even when shown production-ready prototypes, would not add them to their product lines. They wanted cheaper and more flexible. Or at least that was the frustration the ODMs expressed.

From a software perspective, I began to understand why PC laptops were like they were. For example, they were relatively larger than Apple laptops to enable component swapping as though they were desktops. Nearly every review of a Windows laptop bemoaned the quality of the trackpads—both the hardware and software—and here again the lack of focus on end-to-end, including a lack of unified software support from Windows and a requirement for multi-vendor support, made delivering a great customer experience nearly impossible. It was also the case for cellular modems, integrated cameras, Bluetooth, and more.

The desire for lower costs would preclude advanced engineering and innovation in cooling. This led to most Windows laptops to have relatively generic fans and overly generous case dimensions to guarantee airflow and cooling for a flexibility in componentry. The plastic cases filled with holes and grills were just the downstream effect of these upstream choices. So was the fan noise and hot wind blowing out the side of a Windows laptop.

Review after review said Windows laptops didn’t compare to the leading laptops from Apple, and yet I could see the potential to build them if only one of the OEMs would buy them. They simply didn’t see the business case. Their view was that the PC was an extremely price-sensitive market, and their margins were razor thin. They were right. Still, this did not satisfactorily explain Apple nor the inability to at least offer a well-made and competitive Windows PC. Though simply offering it would prove to be futile because of the price and limited distribution such a low-volume PC would necessarily command. We were caught in a bad feedback loop.

A larger initiative Roanne and team would take on was the infamous “crapware” issue, the phrase coined by Walt Mossberg years earlier. Tech enthusiasts and also reviewers tend to view crapware through that namesake lens as software that just makes the PC worse. OEMs had a decidedly different view and to be honest one that as I learned about it, I worked hard to be more sympathetic to. To the OEMs, additional software was a means of differentiation and a way to make their own hardware shine. We tended to think of crapware as trial versions of random programs or antivirus software, but to the OEMs this was a carefully curated set of products they devoted enormous resources to offering. While some were trials and services designed to have revenue upside, others were developed in-house and were there because of unique hardware needs. The grandest example came with Lenovo ThinkPads called ThinkVantage Technologies or TVT. Under this umbrella the full enterprise management and control of the PC hardware was enabled by Lenovo. While I might have an opinion on quality or utility, there was no doubt they were putting a good deal of effort into this work and most of the features were not part of Windows.

Previously I described the OEM relationship as adversarial. That might have been an understatement. The tension with OEMs was rooted in a desire to customize Windows so that it was unique for each OEM or each product line from an OEM. Microsoft’s view was these customizations usually involved “crapware” and that Windows needed to remain consistent for every user regardless of what kind of PC it ran on. It is easy to see this is unsolvable when framed this way. It was amazing considering how completely and utterly dependent Microsoft was on OEMs and OEMs were on Microsoft. This is a case where new people with a new set of eyes had a huge opportunity to reset the relationship.

What I just described was a few months of my own journey to understanding why PCs were like they were, and now I felt I understood. Because I was new, and we had a new team led by Mike, we were optimistic that we could improve the situation. There was no reason to doubt that we could. We believed we understood the issues, levers, and players. It seemed entirely doable. We just had to build trust.

The primary interaction with OEMs were “asks” followed by “request denied” which then led to a series of escalations and some compromise that made neither happy. This would be repeated dozens of times for each OEM for many of the same issues and some specific to an OEM. As with any dynamic where requests are denied, the result was an ever-increasing number of asks and within those an over-ask in the hopes of reaching some midpoint. If an OEM wanted to add four items to the Start Menu, then the ask might be to “customize every entry on the Start Menu” and a reply might come back as “none” or “no time to implement that” and then eventually some compromise. It was painful and non-productive.

There are two schools of thought on these issues. One is that it is obvious that Windows is a product sold by Microsoft and the OEMs should just pass it along as Microsoft intended it to be—relegating the OEM to a wholesale distributor. In fact, the contract to buy Windows from Microsoft specifically states that the product is sold a certain way that should remain. Given the investment OEMs made in building a PC, they did not see the situation that way at all.

The other is that it is equally obvious that the OEMs are paying Microsoft a huge amount of money and they own the customer relationship, including support, so they should be able to modify the product on behalf of their end-user customers. As it would turn out, the first view was on pretty firm ground, at least prior to the 1990s antitrust case. After settling that case, the conclusion was that the right to modify the product as an end-user would pass through to OEMs and Microsoft had limits in what it could require. (If people reading this are thinking about how Android works with phone makers today, you have stumbled upon the exact same issue.)

Perhaps using the word adversary as I have previously used was too blunt. In fact, the relationships were often far more complex and nuanced, fraught with combinations of aligned and mis-aligned incentives. It was entirely the case that the OEMs were our customers, but that was confusing relative to end-users who were certain they were buying a Microsoft product while also a PC maker’s product, and often the Microsoft brand was paramount in the eyes of consumers. The OEMs were also Microsoft partners in product development—we spent enormous sums of money, time, and resources to co-develop technologies and the whole product—yet this partnership felt somewhat one-way to both sides. OEMs were often viewed as competitors as they ventured into Linux desktops and servers, while at the same time they viewed Microsoft as offering one of several choices they had for operating systems. We certainly viewed OEMs as the source our shortcomings relative to Apple hardware, yet the OEMs viewed us as not delivering on software to compete with Apple. Even as distribution partners, depending on who was asked, the OEMs were distributors of Windows or Windows was a distribution tool for the PC itself.

These dysfunctions or as most schooled in how the technology stack evolved referred to them as “natural tensions” had been there for years. There’s little doubt that the antitrust settlement essentially formalized or even froze the relationship, making progress on any part a challenge. There was no doubt a looming threat of further scrutiny as the ongoing settlement-mandated oversight body remained. I met regularly with the “Technical Committee” and heard immediately of anything that might be concerning. Every single issue was resolved, and frankly most trivially so, but the path to escalate was always there.

At the conclusion of the antitrust case, the Final Judgement or Consent Decree (CD) formalized a number of aspects of the relationship, some surprisingly made it more difficult to produce good computers and others simply reduced the flexibility the collective ecosystem had to introduce more competitive or profitable products. While the CD was generally believed to focus on the distribution of web browsers and Java (or in other markets, media players and messaging applications) it also dictated many of the terms of the business licensing relationship. Many of these related to the terms and conditions of how Windows could be configured by OEMs, essentially extending the OEM rights beyond simply installing competitive browsers to just about any software deemed “middleware” as Java was viewed.

In terms of bringing Windows PCs to market, the CD had three main sections:

* Non-retaliation. The first element was that Microsoft could not retaliate against OEMs for shipping software (or middleware) that competes with Microsoft, including shipping computers running competitive operating systems. This codified the right of OEMs to ship Linux and even dual-boot with Windows for example. My how the world has changed, now that is a feature from Microsoft!

* Uniform license terms. Basically, this term meant all OEMs were to be offered the same license and the same terms. Historically, as most all businesses see with high-volume customers, there were all sorts of discounts and marketing programs that encouraged certain behaviors or discouraged others. Now these offers had to be the same for all OEMs (for the top 10 there could be one set and then another for everyone else). In many ways, the largest OEMs got what they wanted but not since this put the best customer on the same playing field as the tenth best. This did not preclude the ongoing Windows Logo program described below.

* OEM rights. The CD specifically permitted a series of actions the OEMs could take with regard to Windows including “Installing, and displaying icons, shortcuts, or menu entries for, any Non-Microsoft Middleware or any product or service” and “Launching automatically, at the conclusion of the initial boot sequence or subsequent boot sequences, or upon connections to or disconnections from the Internet, any Non-Microsoft Middleware.” These terms, and others, got to the heart of many of the tensions with OEMs over “crapware” or software that many consumers, especially techies, complained about. Now, however, it was a right OEMs had and Microsoft could do little to enforce that.

There were many other terms of course. Since we were “stuck” with each other, with Windows 7 we tried a nearly complete reset of the OEM relationship.

The bulk of this work was captured in the Windows logo program—these were the instructions from Microsoft for how OEMs installed and configured Windows during manufacturing described in the OEM Preinstallation Kit (OPK) authored by Microsoft with ample compliance scrutiny. Given the above, one could imagine that most anything was permitted. However, Microsoft retained the rights to implement a discount program for strictly adhering to a set of constraints to earn the “Designed for” logo. Sometimes these were discounts and other times marketing dollars for demand generation, though these are equivalent and interchangeable pricing actions by and large.

We viewed these constraints as setting up a PC to be better for consumers. OEMs might agree, but they also saw them as hoops to jump through in order to get the discount, which their margins essentially required. These logo requirements ran the gamut from basics like signing device drivers and distributing available service packs and patches to how much to configure the Start Menu. None of them ever violated the CD as per ongoing oversight. It was these aspects of Vista that were particularly adversarial.

We revisited all the terms and conditions in the OEM license (and logo) and worked to make the entire approach more civil. We wanted OEMs to have more points of customization while also building the product to more robustly handle those customizations—we aimed to reduce the surface area where OEMs could fracture the Windows experience. We also would spend a lot of energy making the case for why not changing things would be a good idea and in favor of lower support costs or more satisfied customers. A key change we made from the previous structure was to have the same team responsible for taking feedback and producing the logo requirements. This eliminated the organizational seam as well as a place for OEMs to see potential conflicts across Microsoft, and exploit those.

As an example, early in the process of building Windows 7 Roanne and team would present to the OEMs examples of where they were given rights (and technical capabilities) to customize the out of box experience (OOBE) when customers first experienced a new PC. In one iteration we showed a color-coded screens where the OEM customizable region was clearly marked indicating just how much of Windows was open to OEM customization. This type of effort was a huge hit.

The process of collaborating with the OEMs was modeled after our very early Office Advisory Council (OAC) that was run out of the same product planning group where MikeAng and Roanne were previously. Instead of just endless slide decks that spoke at the OEMs, we engaged in a participatory design process with the OEMs, a planning process. We listened instead of talked. We gathered feedback. Then we answered their specific questions.

We began the process of working with the OEMs just as we completed the Vision for Windows 7. We did that because by then we had a real grasp of what we would deliver and when, and at least what we said was rooted in a full execution plan. This was decidedly different than past engagements where the meetings with OEMs reflected the historic Windows development process, which meant that much of what was said at any given time could change in both features and timing. For the OEMs with very tight manufacturing schedules and thin margins, information that was so unreliable was extremely costly to them. At every interaction they had to take the information and decide to act on it, allocate resources, prioritize a project, and more, taking on the risk that the effort would be wasted. Or they could choose to delay acting on some information, only to find out it was critical to start work to provide feedback that might be significant. It was a mess.

What Microsoft had not really internalized was just how much churn and the level of real dollars it cost the OEMs with this traditional interaction. In fact, most involved on the Microsoft side felt, as I would learn, quite good about the openness and information being shared. I learned this both from the OEM sales team and directly from the CEOs of OEMs who were starting to get rather uncomfortable with the lack of information. They began to get nervous over my lack of communication (literally me and not the OEM team) as I had not provided details of the next release even after being on the job for months. I just didn’t know the answers yet. Our desire to be calm and rational created a gap in communication that worried people. It was not normal. We had not set expectations and even if we had it wasn’t clear they would have believed us.

MikeAng characterized the old interaction as “telling the OEMs what we were thinking” when we needed to be “telling OEMs what we were doing.” By characterizing what we were going to improve this way we were able to avoid massive amounts of wasted effort and negative feelings. We also believed we would reduce the number of budget-like games when it came to requests. No matter how much Microsoft might caveat a presentation as “early thoughts” or “brainstorming,” customers looking to make their own plans under a deadline will hear thoughts as varying degrees of plans. Only later when those plans fail to materialize does the gap between expectations and reality widen—the true origin of dissatisfaction. At that point the ability to remind a customer of whatever disclaimer was used is of little value. The customer was disappointed. This cycle repeated multiple times over the 5-year Vista product cycle, and many before that over the years.

With the Windows 7 vision in place, we began a series of OEM forums and meeting at least every month with each OEM. Each forum (an in-person workshop style meeting) would focus on different parts of the product vision or details about bringing Windows 7 to market. Roanne and team dutifully documented the feedback and interactions. We actively solicited feedback on priorities and reactions to the product overall. Then we summarized that work and sent out summary of engagement memos to the OEMs. In effect, the relationship was far more systematic, and the information provided was far more actionable. From the very start we communicated a ship date for the product and milestones, and a great deal of information about the development process. We reinforced our attention to the schedule with updates and information along the journey.

The entire process was run in parallel for the key hardware makers: storage, display, networking, and so on. We also repeated it for the ODMs. The team did an amazing job in a very short time, for the first time.

Over the coming months and really years, the reestablishing of trust and the effort to become a much more reliable, predictable, and trustworthy partner would in many ways not only change this dynamic and experience for OEM customers, but significantly improve the outcome for PC buyers. At the same time, OEMs were able to see improvements in satisfaction and potential avenues to improve the business. The PC would see many major architectural changes over the course of building Windows 7—the introduction of ink and touch panels, broad use of security chips and fingerprint login, addition of sensors, expansion of Bluetooth, transition to HDMI and multiple monitors, high-resolution panel displays, ever-increasing use of solid-state disks, and the transition to 64-bits. The Ecosystem team would be the conduit and moderating force between OEMs, their engineering teams, the engineering teams on Windows, and even Microsoft’s own OEM sales and support team.

The Windows business faced a lack of trust from every business partner. The Windows team needed to move from chaotic process that promised too much and delivered inconsistently and late. Instead, we aspired to make bold but reliable promises—my guiding principle, the mantra of promise and deliver. While we talked big, the ball was in our court. We put in place a smoother and more productive engagement with OEMs. Throughout the course of developing Windows 7 and beyond, we consistently measured the success of the engagement with qualitative and quantitative surveys. Through that we could track the ongoing improvement in the relationships.

Promise and deliver.

Just as we worked to gain renewed focus with OEMs, Apple chose to declare war with Microsoft, and the PC, with clever and painfully true primetime television commercials. While the commercials began before Vista, the release and subsequent reception of Vista were exactly the material the writers needed for the campaign to take off.

On to 090. I’m A Mac