Data vs. Commerce

For eight months of trade shows, AI talk meant guardrails. At B2B eCommerce World Americas in Indianapolis, practitioners were talking about how they’d use it. They were also talking about the data quality problems that made them backtrack.

ㅤ

Floyd Blaikie missed the trip, so she gets the debrief from Matt Johnson and two members of the Pivotree team, Ryan Smith and Matt Fowler. Both guests took the data side.

ㅤ

Ryan met practitioners who had hired more than three SIs that never handled the data part of the work. Matt Fowler heard marketing, ecom and executive leadership calling the boring data side of it a company challenge. They also trade notes on the AI matchmaking app and the badges with no company names. Then each of them predicts what next year’s show will be about.

ㅤ

👤 Guest Bios

Ryan Smith is an Account Executive on Pivotree’s Industrial Manufacturing and Distribution team, working with distributors and manufacturers on AI and data. This is his second time on Data vs. Commerce after Episode 7. He took the data side, arguing that the companies doing it right make somebody own the data internally.

ㅤ

Matt Fowler is an Account Executive at Pivotree covering automotive and transportation. He has spent more than a decade in automotive aftermarket digital work, including 360-degree product imaging standards. He won a 2026 Auto Care Association Impact Award. He took the data side here, pointing out that data has stopped being something IT plugs in and makes work.

ㅤ

📌 What We Cover

• AI conversations at the show moved from guardrails to “we’ve gotta get into this yesterday.”
• Teams that started their digital transformation had to backtrack because the data wasn’t where it needed to be in terms of quality.
• Some practitioners hired more than three SIs, and none of those SIs handled the data part of it.
• Teams are stepping back from build versus buy after finding re-platforming and data governance harder than planned.
• The number one problem in organizations is still people, governance, process and buy-in.
• The companies doing it right make somebody own the data internally.
• The predictions for next year are change management, tribal knowledge leaving with retirees, and AI becoming an assumption.

ㅤ

🔗 Resources Mentioned

• B2B eCommerce World Americas 2026, run by the B2B eCommerce Association (B2BEA)
• Akeneo
• OroCommerce
• Inriver
• Stibo Systems STEP
• Pivotree Intelligent Managed Services

Mike Dattilo described his current integration project as a jigsaw puzzle shaken out on the table. Floyd Blaikie corrected him to two or three puzzles, and he agreed.

ㅤ

Mike runs eBusiness at Xerox, covering B2C, B2B and partner commerce. He’s folding an acquired organization’s systems into Xerox infrastructure that wasn’t fully integrated to begin with, which adds up to literally dozens of systems.

ㅤ

Mike took the commerce side with Matt Johnson and Floyd. The friction he kept returning to was human: a pricing spreadsheet John loads every day that never came up in requirements, and an OST report that surfaced only after go-live.

ㅤ

He walks through the documentation his team keeps and why they run Kanban instead of a two-year waterfall plan. He also covers what he’d do the moment an acquisition is announced. Data vs. Commerce is produced by Pivotree.

ㅤ

👤 Guest Bio

Mike Dattilo is Director of eBusiness at Xerox, where he leads direct commerce across B2B, B2C and partner channels. He came up through Lexmark, starting as a software engineer and later serving as Digital Marketing Director, where he led Lexmark’s move to Adobe Commerce on Adobe’s cloud.

ㅤ

Xerox acquired Lexmark in July 2025, so Mike is now integrating systems he used to run. He co-organizes the Adobe Commerce Virtual USA User Group. On this episode he took the commerce side.

ㅤ

📌 What We Cover

• Architecture documentation at both extremes: a Word doc map that looked drawn in Microsoft Paint, and AI-generated docs hundreds of pages long.
• Why requirements gathering is usually one of the longest cycles, and why leadership asks for implementation time before the team knows what it’s implementing.
• Mike’s working set: a one-page architecture diagram, data flow diagrams that map where PII resides, generated code documentation, and automated unit tests that catch a broken order confirmation email before a customer does.
• Critical singularities, from John’s daily pricing spreadsheet to the OST report nobody on the team had heard of.
• Building cross-training time into the schedule when it doesn’t generate new revenue.
• Running Kanban instead of a two-year waterfall plan, and why ten top priorities means equally slow progress on all ten.
• What to do the moment you hear you’re being acquired, and choosing what’s best for the company right now over the best available software.

ㅤ

🔗 Resources Mentioned

• Adobe Commerce
• SAP S/4HANA
• Smartsheet
• Microsoft Word and Microsoft Paint
• Adobe Commerce Virtual USA User Group (Mike’s quarterly user group)
• Pivotree integration services (episode sponsor)

A maintenance manager at a food plant orders a replacement for a sheared conveyor bolt. It shows up two days later, looks perfect, and won’t go in the hole. For the Halloween episode, Matt Johnson and Floyd Blaikie of Pivotree skip the interview and play product data telephone as a radio drama.

ㅤ

One fictional M10 hex bolt travels through five handoffs. It goes from a manufacturer spec sheet to item setup at a distributor, then to a marketplace listing and an inside sales call. From there it reaches a purchasing system and finally an AI purchasing agent with permission to buy equivalents. Each person only hears what the person before them said.

ㅤ

Matt and Floyd track eight attributes: size, length, thread pitch, material, thread length, standard, head size, and how many come in the box. Two make it through. Floyd owns one of the deaths herself, since the listing writer’s job is her job. Every company and character is made up, and every voice besides the hosts’ is AI generated.

ㅤ

📌 What We Cover

• Dana’s ERP template, which has a 40-character short description and a material dropdown limited to stainless, carbon, or zinc plated, with no spot for a grade.
• The description graveyard: the long description field that catches everything the template didn’t plan for, which nothing downstream ever reads.
• Tyler deciding PT means premium thread, then adding grade 8 because every top seller in the category says it, and why the engineer’s words on a spec sheet don’t rank.
• Ray asking the right questions twice, about the thread and the chlorine washdown, and Marcus answering from the only screen he had open with “M10’s M10.”
• Fastenal’s third quarter, where more than 61% of sales were digital, much of it through bins and vending machines on customers’ floors that already know the part number.
• Linda’s “allow equivalent substitutes” checkbox, ticked because finance wants three quotes on anything under five hundred dollars.
• The AI purchasing agent that matched every attribute it was given, bought coarse thread A2-70 in a pack of 100, and reported $11.40 in savings.
• Where the chain could have been saved: a field for thread pitch and a grade in the dropdown at the first handoff.

ㅤ

🔗 Resources Mentioned

• Fastenal, third-quarter results
• DoorDash, agentic ordering for offices and corporate customers
• Walmart, agent profiles with spending caps
• DIN 960 and ISO 4017 fastener standards
• B2B eCommerce World, Indianapolis, November 2 (Pivotree at Table 57)
• B2B Online Atlanta, November 9 to 11, with a live Data vs. Commerce episode (Pivotree at Booth 306)

Floyd Blaikie got so mad at every reporting tool on the market that she vibe coded her own. That put her inside a Gartner prediction: 60% of the vibe-coded apps running in businesses today will be gone by 2028. She and Matt Johnson brought in two Pivotree colleagues to argue about which of those apps deserve to stick around.

ㅤ

Joel Farquhar, Pivotree’s CTO, thinks the 60% is low. Software has always been disposable, he says, and not a single piece of software he ever wrote is still in production. Aiden Parker, a senior product manager on Pivotree’s data management suite, described both ends of what commerce teams are building: SEO copy workflows, and a film and entertainment equipment distributor that built its own order management and tracking tools.

ㅤ

Neither guest took the data or commerce side. The friction ran between Floyd, who argued that vibe coding skips the part where people sit down and ask whether something should be built, and Joel, who’d rather learn an idea is garbage in a day than in four months. Joel walked through Pivotree’s internal experiment registry. He ended on the maintenance problem with a line he borrowed: “everybody likes to vibe code, but nobody wants to vibe maintain.”

ㅤ

👤 Guest Bios

Joel Farquhar became Pivotree’s Chief Technology Officer on July 1, 2026, after serving as Chief Architect, and reports to CEO Bill Di Nardo. He’s been with Pivotree for about 20 years. He spent the first 15 years of his career as a software developer, and his work has centered on enterprise ecommerce for retailers, distributors, and manufacturers. He studied computer science at Carleton University and is based in Ottawa.

ㅤ

Aiden Parker is a Senior Product Manager at Pivotree, where he’s spent close to two years on the data management suite of products. Before that, he was Head of Product at Fable Tech Labs and held product roles at Flashfood, Inkbox, and Wattpad. He’s tried out senior vibe coder and citizen developer as titles, but he’s going with product builder for now. Anything with the word developer in it feels disingenuous to him.

ㅤ

📌 What We Cover

• Why Joel thinks Gartner’s 60% figure is low, and why “60% of what?” matters when the count runs into the thousands
• Both ends of what commerce teams are building: SEO and product copy workflows, and a film and entertainment equipment distributor building its own order management, delivery, and tracking tools
• Why the vibe-coded tools that last are small operational efficiency apps that move data between parts of an organization, while one-off dashboards get abandoned
• What Matt saw at Applied AI for Distributors: an explosion of order automation software, and developers studying it so they could code their own
• Why most of these apps never leave the creator’s desktop, and why adoption is harder than the build
• Floyd’s case that vibe coding skips the conversation about whether to build, and Joel’s case that finding out in a day beats finding out four months later
• How Pivotree’s internal experiment registry worked, including AI-flagged duplicates and a certification layer for anything being promoted
• Who keeps a vibe-coded app running 24/7 once the business depends on it

ㅤ

🔗 Resources Mentioned

• Gartner research on governing vibe coding for citizen developers, the source of the 60% by 2028 prediction
• Claude Code
• Applied AI for Distributors, the Distribution Strategy Group event Matt attended
• Jurassic Park and Dr. Ian Malcolm’s line about being so preoccupied with whether you could
• Pivotree Intelligent Managed Services for Stibo STEP customers, including the free Intelligent Health Check

It’s Q4, budgets are locking, and every data and commerce request has to be translated into what a distributor’s or manufacturer’s board cares about: growing existing accounts, winning new logos and protecting margin. Matt Johnson and Floyd Blaikie of Pivotree take this one solo, working through the questions boards ask during strategic planning and the numbers a practitioner should answer with.

ㅤ

Matt draws on his time as a go-to-market consultant in distributor and manufacturer boardrooms. He starts share of wallet with historical data and digital adoption, ties new-customer acquisition to product data and visible pricing, and argues the traditional sales rep role is going away or changing. Floyd presses him on the metrics and on what stays with reps. With no guest assigned a side, the friction lands between commerce and sales.

ㅤ

📌 What We Cover

• Digital adoption as the first B2B ecommerce number for the board: portal use, quoting tools, one-click repeat orders, and retention of digital versus offline accounts
• How inconsistent or incomplete product information keeps a distributor out of SEO and AEO results
• Why pricing and inventory behind a login or a call for quote won’t convert a new customer
• Channel conflict when a manufacturer sells direct, and Matt’s workaround: a high list price, a 25% first-time discount, then account-based onboarding to contract pricing
• Cost to serve for a self-serve order versus one that needs five human touchpoints, and Matt’s estimate that the portal is sometimes 80 to 90% less expensive
• Why returns that aren’t digital mean at least five phone calls
• Sales and commerce data that doesn’t sync: online orders missing from the CRM, offline orders missing from the portal
• An AI plan for 2027 built around the top five areas where manual workarounds cost money

ㅤ

🔗 Resources Mentioned

• Sam Schwartz, Founder, Sierra Summit Digital, on Ep. 17: The backwards search from spec to SKU
• Stibo Intelligent Managed Services from Pivotree for Stibo STEP customers (episode sponsor)
• Stibo Intelligent Health Check (free)

Agentic commerce got pitched as a personal shopper that takes your request and spends your money without checking back. What shipped is a discovery tool that puts a shortlist on your screen and waits for you to approve the purchase.

ㅤ

Matt Johnson and Floyd Blaikie of Pivotree sat down with Paul Drecksler, founder and editor of the Shopifreaks newsletter, to agree on a definition. Paul took the commerce side. His definition is the strict one: the whole transaction runs autonomously, from discovery to purchase, without re-engaging the user.

ㅤ

Then he explained why no live tool meets it. Chargeback and return liability sits with the merchant, courts are treating agents as tools rather than people, and shoppers enjoy the treasure hunt. Floyd asked who gets in trouble when the wrong dishwasher shows up. Matt made the case for B2B replenishment leading the way, and Paul pushed back that B2B platforms are decades behind retail.

ㅤ

👤 Guest Bio

Paul Drecksler founded Shopifreaks in January 2021. It's a weekly newsletter that recaps e-commerce, B2B, D2C and advertising news for merchants and operators, and it's closing in on its 300th edition. He reads thousands of articles a week so subscribers don't have to. He also consults through Ideas Focused, which includes B2B client work, and he's a longtime traveler who has lived and worked across Latin America. He took the commerce side.

ㅤ

📌 What We Cover

• Paul's strict definition of agentic commerce: discovery, product choice and payment all happen autonomously, based on the user's first input.
• Why discovery today is "glorified search," and why embedded one-click checkout isn't new.
• Amazon's set-your-price auto-buy and Floyd's Subscribe & Save test: are these pre-approved clicks or agentic commerce?
• The B2B case: bills of materials and replenishment suit agents, but catalogs still live in sales reps' heads, PDFs and spreadsheets.
• Who absorbs the cost when an agent buys the wrong dishwasher, and why chargebacks and returns land on the merchant.
• Amazon vs. Perplexity, and the legal definition forming around agents as tools.
• Why Paul still does his own follow-up research when output accuracy sits below 99.9.
• Paul's mom, a Facebook group colorizing her photos, and what mainstream adoption looks like.

ㅤ

🔗 Resources Mentioned

• Shopifreaks
• Pivotree integration services
• Heather Hershey, IDC
• Zack Kass
• Google's Universal Commerce Protocol
• ChatGPT Instant Checkout
• Amazon Alexa and Subscribe & Save
• Amazon v. Perplexity
• Meta's Muse shopping agent
• T.J. Maxx

Distributors think their customers want reps who know the catalog. Their customers mostly want to know where their stuff is.

ㅤ

Floyd Blaikie brings Matt Johnson a report she got lost in the morning they recorded. It's the State of Sales in Distribution 2026 from Phocas, built on a survey of 100 distributors. Floyd reads the numbers. Matt, who leads go-to-market for distribution at Pivotree, explains why they look the way they do.

ㅤ

This is a hosts-only episode, so nobody takes a side. The friction comes from the data itself. Then Matt admits that his own sales agent hallucinated on unmaintained data, and his step one for distributors has nothing to do with buying AI.

ㅤ

📌 What We Cover

• Why 34% of distributors don't measure new business win rate, and 39% don't track or don't know their forecast accuracy
• Why the data used by the most people inside a distributor gets dirty the fastest
• 75% of distributors prize rep product knowledge, while about 5% say their customers want technical expertise
• What happens when product knowledge lives in reps' heads instead of the product database
• Why electrical buyers care about fast quotes and "where's my order?" more than discovery
• 55% of distributors call field sales their top channel and 5% say e-commerce, which Matt reads as being in the retention business more than the growth business
• Every distributor with no plans to adopt AI has a sales team under 20 reps
• Matt's step one: get the people who maintain the data into the strategic meetings

ㅤ

🔗 Resources Mentioned

• State of Sales in Distribution 2026, from Phocas
• Data vs. Commerce Episode 6 with Tromml
• Amazon Business
• Pivotree

Sell sheets, spec sheets and price books are the collateral B2B salespeople say they can't work without, and they go out of date the moment the catalog changes. Some manufacturers can only refresh a price book every 12 to 18 months.

ㅤ

Matt Johnson and Floyd Blaikie of Pivotree sat down with David Feinleib, founder and CEO of It'sRapid, which generates marketing collateral straight from PIM data. David took the data side. His argument: when you produce thousands of documents from product data, every missing spec or diagram turns into a gap on the page. One customer told him it felt like an audit for their data.

ㅤ

Matt pushed on governance and why stale data gets shipped anyway. Floyd asked whether the gaps turn into finger pointing. David recast the fight as data versus design, the left brain and right brain of every sell sheet.

ㅤ

👤 Guest Bio

David Feinleib is founder and CEO of It'sRapid, a Palo Alto company that turns product data from PIM and DAM systems into sell sheets, price books, product images and retail media ads. Before It'sRapid he founded Content Analytics, an ecommerce analytics and product content platform that Syndigo acquired in 2019. He's a former general partner at Mohr Davidow Ventures, led product management teams at Microsoft, and wrote Why Startups Fail. He took the data side on this episode.

ㅤ

📌 What We Cover

• Why a customer told David, "I didn't realize I bought an audit tool," and what the gaps on the page revealed about their product data
• Price books that can only be updated every 12 to 18 months because the data keeps changing on the team
• The legacy workflow: custom scripts pushing data into templates, then sign off from marketing and legal
• Regenerating 10,000 pieces of collateral after a logo change, and what happens when one field flows through to 10,000 sheets
• Matt's case that the problem often happens in governance: letting stale data in and shipping it anyway
• Webhooks as the ideal state, where a data change triggers regenerated sheets automatically
• Data and design as David's version of data vs. commerce, and the client who went from one region to seven or eight
• Whether print is dead, tested against a printed birthday card versus an e-card

ㅤ

🔗 Resources Mentioned

• It'sRapid (David's email: [email protected])
• Pivotree Intelligent Managed Services for Stibo STEP customers, including the free Intelligent Health Check
• Stibo Systems STEP
• Data vs. Commerce Episode 6 (field sales and AI)

Increasing share of wallet among existing customers is a huge priority for manufacturers and distributors. On paper it looks easy. In practice, marketing blames sales for not selling more into the base, sales blames marketing, and both blame IT for not prioritizing their stuff.

ㅤ

Floyd Blaikie put a third suspect on the table: maybe neither is the problem, and the problem is actually data. She and Matt Johnson of Pivotree took that to Sam Schwartz, who took the data side and stayed there. Her position: the data is accessible. It's probably in your ERP somewhere, probably in a catalog somewhere. What's hard is prioritizing it, and that comes down to fear and ego at the leadership level.

ㅤ

The conversation gets into project-based quoting, the backwards search from a spec to a SKU, the order desk as the place where AI is landing hardest in B2B, and how to build a product data business case an exec team will fund.

ㅤ

👤 Guest Bio

Sam Schwartz has spent about 17 years working for distributors and manufacturers on product data, digital transformation, and technology adoption programs, including senior digital roles at Jensen Precast and Midland Industries. She now runs her own consultancy, Sierra Summit Digital, working with mid-market B2B companies, and she is Global Director of Practitioner Experience at the B2B eCommerce Association. She took the data side.

ㅤ

📌 What We Cover

• Why siloed thinking between sales, marketing, and IT stalls customer expansion before data ever enters the conversation
• The sales team that had decent product data and still couldn't cross-sell, because the only thing in their tools was what the customer bought last time
• Why the real blocker is prioritization, not access: the data is probably already in the ERP or a catalog
• Fear and ego trickling down from leadership teams, and what accountability looks like when a CEO's vision gets ignored to his face
• Cross-category and consumables campaigns, and where they work in a broad-line catalog
• The backwards search: bids that arrive as "2,700 units of a three-inch steel pipe" instead of part numbers, on lists thousands of lines long, and quotes that take weeks
• Why the order desk is where AI is making the biggest impact in B2B, and why it depends entirely on structured product data
• How to build the business case without asking for a million dollars: quotes won, speed to quote, cost to acquire, and sales team time

ㅤ

🔗 Resources Mentioned

• Sam Schwartz on LinkedIn (she responds to messages quickly)
• Sierra Summit Digital, Sam's consultancy for mid-market B2B companies
• B2B eCommerce Association
• B2B eCommerce World, Americas event in Indianapolis, November second and thirdㅤ
• Pivotree integration services: integrations in half the time and half the cost, monitored in production for a fixed monthly fee, starting with a free integration sprint

A packaging change is a data change, and most brands don't treat it as one. An item profile gets built months in advance, pitched to a retailer as a 12-pack, and then a new channel wants six or eight to a box. Nobody updates the profile. The POs arrive at the warehouse and the unit measures don't marry.

ㅤ

Matt Johnson and Floyd Blaikie sit down with Leo Rodriguez, VP at River Plate, Inc., a 3PL in Southern California. Leo took the data side, which is a more interesting position than it sounds coming from the person whose team physically touches the box.

ㅤ

Floyd opens by reading a line from his LinkedIn back to Matt: the operation is only as clean as the data behind it. What follows is what that looks like at the end of the chain. Bad mapping that shows a green check mark. ASNs going out wrong for two weeks. Chargebacks that surface six months later, when corrective action is no longer an option.

ㅤ

👤 Guest Bio

Leo Rodriguez is Vice President at River Plate, Inc., a family-owned third-party logistics provider in Simi Valley, California. The business started as local trucking and freight in Los Angeles and now runs a 138,000 square foot facility covering e-commerce fulfillment, wholesale and retail distribution, value-added services, cross-docking, and freight logistics. Leo has been there 17 years, going into his eighteenth. He joined without a supply chain background, took courses online, and learned the operation on the floor.

ㅤ

📌 What We Cover

• Why the item profile and the purchase order stop matching, and why the 3PL is the last to know
• What happens when a brand sets up a 12-pack, then sells the same SKU as a six-pack on a different channel
• Why a green check mark on an integration doesn't mean the data arrived correctly downstream
• The mapping change on a tech provider's back end that nobody communicated until Target sent an email
• Chargebacks that don't surface until six months after the mistake, and what that does to corrective action
• The onboarding red flag: five bullet points sent to a brand, two of them answered
• Pre-mapped connections to Walmart Marketplace and Shopify versus a REST API and what each one lets you control
• The reverse logistics client who built his own returns dashboard with Claude and connected it in 15 minutes
• Leo's advice: test the reality, not the PowerPoint, and go past go live before you believe it works

ㅤ

🔗 Resources Mentionedㅤ

• River Plate, Inc. (RFQ page and discovery call)
• Walmart Marketplace
• Shopify
• Claude
• Target
• Kohl's
• Amazon
• Salesforce