I’ve spent my career trying to marry art and science. The creative and the analytical. The gut instinct and the spreadsheet. Never leaning too far in either direction because the best decisions live in that tension.

So when Philip Armstrong told me most executives use “we need better data” as a stalling tactic, I immediately recognized the pattern. I’ve seen it a hundred times. Hell, I’ve probably done it myself.

It’s not that they don’t trust the data. It’s that they don’t trust themselves to be wrong. No one ever got fired for hiring IBM.

Philip Armstrong is ex-Nike, now SVP at Truth and Data. He’s rebuilt customer journeys inside some of the world’s best brands and messiest data orgs. And he’s seen the same pattern everywhere: companies drowning in data, starving for action.

Here’s what he told me that I can’t stop thinking about.

There Is No Perfect Data. Stop Pretending There Will Be.

“I see a lot of people waiting for the perfect data to make decisions,” Philip said. “And there’s no scenario where you have perfect data.”

Let me repeat that for the leaders hiding behind their dashboards: there is no scenario where you have perfect data.

Decision-making with incomplete information isn’t a bug. It’s the job. If you had perfect information, you wouldn’t need judgment. You’d need a spreadsheet and an intern.

But “we need more research” is socially acceptable. It sounds responsible. Strategic. Thoughtful.

It’s actually just expensive cowardice.

Your competitors aren’t waiting for perfect attribution. They’re making the call with the best data they have right now and iterating based on what they learn.

You? You’re waiting for a certainty that will never arrive.

Thanks for reading Playbook Broken! Subscribe to get it in your inbox!

Three Conversations Beat Three Thousand Data Points

Here’s Philip’s move that sounds insane until you think about it:

“When you interview like three or four people, you get 90% of the most significant changes that you need to do. If you interview another thousand people, you get a long tail list of stuff that doesn’t really move the needle.”

He learned while doing user experience testing for streaming services. Three interviews led to picture-in-picture play and sound-off video for social media; features that became industry standard.

When he brought those insights back to leadership, they wanted statistical significance. Sample sizes. They were addicted to the massive web analytics datasets where n=100,000 feels normal. Sound familiar? Does to me.

But here’s what massive sample sizes actually do: they create averages. And averages bury signal.

The interesting stuff lives on the edges. In the outlier behavior. In the thing that three people said that makes you go “wait, what?”

You’ve built an entire reporting infrastructure to avoid talking to customers. You know what device they use, what time they click, how long they scroll.

You just don’t know why they left without buying.

Three phone calls would tell you. Three!

But that requires getting uncomfortable. It requires hearing things that might contradict your roadmap. It requires being wrong in front of people.

Data dashboards let you be wrong in private.

The Nike Lesson. Data Is a Voice, Not a Verdict

Philip spent years at Nike learning something I’ve always believed but never had the right language for: data shouldn’t tell people what to do. It should give them another perspective to consider.

He came from digital functions where metrics drove decisions. Then he landed at one of the most creative companies on the planet, where the brand was built by people who trusted their instincts.

“You encounter people who built the business, who created the edge,” he said. “You show them a table of data and tell them to make a different decision than what their gut tells them. They laugh at you.”

He tried everything to make them more “data-driven.” Better visualizations. More compelling presentations. Nothing worked.

Until he reframed it entirely.

“It was very important not to tell them what the data says, but really present the data as an additional opinion that now sits at your table. Here’s another expert opinion for you to consider.”

Not mandate. Not truth. Just another voice in the room.

“That changed the game quite a bit.”

This is what I mean by art and science working together. The creative leaders at Nike had context Philip couldn’t see in the spreadsheet. They’d talked to the CEO, the competition, manufacturing. They had instincts built from years of making bets.

The data gave them one more input. They didn’t always follow it. But now they were making better decisions because they had both the numbers and the nuance.

That’s the balance. Data as perspective, not prescription. Instinct informed by evidence, not replaced by it.

Your Problem Isn’t Data. It’s Courage.

Most organizations don’t have a data problem. They have an adoption problem disguised as a data problem.

You build dashboards. Nobody acts on them.

You hire analysts. Nobody listens to them.

You buy tools (moar martech!). Nobody changes behavior.

And then you blame the data. “We need better attribution. We need real-time signals. We need AI to surface insights.”

No. You need someone willing to make a call when the data says 75% but feels like 50%.

You need someone willing to override the model when three customer conversations reveal something the algorithm missed.

You need someone willing to say, “We’re moving forward with incomplete information because waiting costs more than being wrong.”

That’s not a data literacy problem. That’s a leadership problem.

And the organizations that win in the next decade won’t be the ones with the best dashboards. They’ll be the ones where leaders make decisions, learn fast, and iterate.

Perfect data is a myth. Perfect decisions don’t exist.

But courageous decisions with good-enough data? Those compound.

If you’ve read this far, why not share this post with someone who might enjoy it?

The Question

If you had to make one decision tomorrow without perfect data, what would it be, and why aren’t you making it today?

Not “what tools do we need?”

Not “what’s our sample size?”

What are you avoiding deciding because you’re scared?

Around the Town

* Holy cow, I’m running a conference… AI for GTM is a one-day experience, in Durham, NC, on December 3rd. As a Playbook Broken reader, I have a free pass for you, just ask.

* Deep into Season 2 planning for Playbook Broken. We’re shifting to more collaborative problem-solving, less traditional interviews. If you’ve got ideas for guests I should talk to or topics that need dismantling, find me on LinkedIn.

* Currently reading Enshittification by my actual favorite author, Cory Doctorow. What have we done, people, what have we done?

* Also: Plur1bus is absolutely bonkers and worth watching. From the people who made Breaking Bad and Better Call Saul.

Thanks for reading. If your playbook is cracked, we’re here to help rethink and rebuild it with you.

—Marc

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit marcsirkin.substack.com