Aran Hamilton put it well recently: "you'd never buy the lumber to build a house before you had the architectural drawings and a general contractor." The rework alone would cost more than doing it right the first time — and that's before you account for living in a half-built house while you sort it out.
This is retail media today. We are living in a half-built house.
I've been in rooms with some of the most sophisticated retail organizations in the world — smart people, serious businesses, real ambition — and almost universally, the technology and operational decisions made in the early stages of building a retail media business follow the same pattern: prioritize the output, defer the foundation. Revenue before scalability.
Today, retail media has most of the pieces it needs. They're just not connected. And with consolidation at the top and slowing growth rates overall, it's time to rethink our approach.
In this week's Retail Leapfrog Series — a collection of thoughts designed to help retailers leapfrog incumbents — I want to talk about organizational debt. The kind that is masked on a P&L by really low COGS and almost no R&D. The kind rooted in manual workflows you'll "wait until next year" to fix. The kind that conflates a one-way door with a two-way door — where manual work becomes the measure of competence, and once engrained, is nearly impossible to break.
Half the nails.
The race for the lowest-cost technology solution here is misguided. If you find yourself saying "there are cheaper solutions out there," it's worth stopping to ask why.
There are always cheaper solutions — I could buy cheaper lumber, tie the house together with fewer nails, hire inexperienced labor — but a like-for-like comparison on square footage misses the point, because one of those houses isn't going to stand up in the wind. Cheaper isn't a discount on the same thing. It's a different thing, priced accordingly.
So it's worth asking why the price is lower in the first place. It usually comes down to one of three reasons:
- The product isn't built yet. You're paying for a roadmap and a demo.
- The vendor isn't investing in R&D. What you're buying is what you're getting; it works today because the market hasn't moved yet.
- Low-cost humans are quietly managing the connections in the background. Manual workflows hiding behind a sleek user interface.
Ultimately, a half-built house isn't built to last. As AdButler put it in a phenomenal byline earlier this year:
The cheap, fast decisions didn't save us anything. They just moved the bill to a moment when it's far more expensive to pay — and that's before you account for everything you'll have built on top of it in the meantime.
One-way doors.
There's the old Amazon adage: one-way vs. two-way door decisions.
With a two-way door decision, you walk through, and if you don't like what's on the other side, you walk back — the only cost of being wrong is the time it took to find out. With a one-way door decision, you walk through and the door closes behind you — walking back is either impossible or expensive enough that you never will.
Yes — making short-term, lower-cost decisions and making up for it with manual workflows or fragmented architecture feels like something you can walk back from. It feels reversible because, in the moment, it genuinely is. Just a workflow. Just a spreadsheet. Just a connection someone is managing by hand until we get around to building the real thing.
But the decision never stays contained to the thing we decided. Over time we build the rest of the business around it:
- The next tactic gets designed to fit the data we already collect rather than the data we actually need.
- The next hire is trained on the process that exists rather than the one we would design from scratch.
- The next integration gets wired into the same connections we told ourselves were temporary.
Sarah Marzano at EMARKETER has pointed out that "most retail media networks were built really quickly and in silos" — and the silo is rarely the thing we go back and fix. It quietly becomes the thing that everything depends on. By the time we decide to walk back through the door, there is an entire business standing on the other side of it.
The hardest part of walking back was never the technology. Andrew Lipsman has framed the underlying tension well — that RMNs "will need to balance meeting short-term revenue goals while making necessary investments to achieve long-term scale" — but the reason that balance is so difficult is rooted in behavior.
The person who stays late to fix the broken report gets promoted; the person who quietly builds a system so the report never breaks doesn't. Organizations learn what they reward, and most retail media organizations have spent years rewarding the firefighter rather than the architect.
New hires inherit the spreadsheet, learn the workarounds, and are recognized for mastering them — so that when someone eventually suggests replacing the whole approach, the response becomes: "we're okay, we can handle this." They aren't raising their hand against a tool that would make their lives infinitely easier. They're challenging how competence has been defined in that organization.
Renovating the house.
Every retailer I talk to about this agrees — spreadsheets are not a long-term answer, and nobody is defending the fragmented technology stack. The plan, almost always, is to fix it later: once the revenue is bigger, once the team is larger, once there is "budget for the rebuild."
It is a reasonable plan. It is also, in my experience, the plan that never actually happens — billion-dollar retail media businesses today are still operating, in part, on spreadsheets or with manual reconciliation.
Forrester research, commissioned by Koddi, found the gap between how mature retailers think they are and how mature they actually are:
Only 12% can seamlessly activate and measure campaigns across on-site, off-site, and in-store. This data also speaks to my earlier point — that inefficiencies have become cultural. 42% believe they're advanced, yet only 12% have actually connected the pipes.
This is also why I keep coming back to Kiri Masters' Retail Media Doom Loop. When there is strong momentum, there is less incentive to innovate or evolve. When the cracks start to show, there is less trust in the retail media organization, and less willingness to invest.
Once growth stalls, the natural instinct is to either invest our way back out of it or find "quick wins" — but the half-built house has quietly taken that option off the table. Building on top of a half-built house is not only cumbersome, it's also less appealing for the buyers (advertisers), ultimately making it harder to buy.
The very thing we were waiting for in order to justify the rebuild — more revenue, more scale, more room to maneuver — is the thing the unbuilt foundation now prevents us from earning. We deferred the work until we could afford it, and in deferring it, we made certain we never could.
I made the financial version of this case in The Economics of Waiting, where every quarter of delay compounds the cost of ever catching up. This is the same argument seen from inside the business rather than from the spreadsheet. The decision is never really between building the foundation now or building it later. It is between building it now and not building it at all.
A house built on purpose.
It would be easy to read all of this as inevitable — the natural cost of building a retail media business under real pressure, with real revenue targets and a narrow window to hit them. But it isn't inevitable, and the clearest evidence I have is a conversation I had on stage recently at the Path to Purchase Institute's Retail Media Summit with Stephenie Cattonar, who leads product for Orange Apron Media by The Home Depot.
What struck me most is that Stephenie didn't start somewhere different from everyone else. She started in exactly the same place. The technology Home Depot was using internally to run its marketing, she told me, "was never designed to support the scale of what becomes the ultimate driver of your retail media network" — the same pressure every retailer feels, and the same systems never built to handle it.
What did she build? A composable stack that connects all consumer touchpoints and all ad tech, unifying every job-to-be-done and workflow into a single platform. This is Orange Access — a managed-service and self-service platform that powers The Home Depot's retail media business today, and eliminates the need to operate it on spreadsheets or over email.
The conviction was hers. We helped her build the house she had already decided to live in. And rather than deferring the hard part until the business forced her hand, she asked it at the beginning:
Those are not the questions a team asks when it's comfortable; they're the questions most organizations only get to once the ceiling is already overhead. Stephenie was asking them on day one — while the business was still in its initial accelerated growth state.
She also did the thing most of us avoid: treat the decision as a matter of evidence rather than instinct. "It was a very data-driven exercise," she said, "understanding the amount of dollars flowing in, the operational burden being put on our associates to execute the media, the profitability associated with that. There is a ceiling, and we were going to hit it fast and furiously if we didn't make different decisions." She could see the ceiling before she reached it — not as a feeling, but as a line on a chart heading somewhere she had no intention of going.
And then she did the thing this entire piece has argued is hardest of all: she changed the foundation while the business was still growing.
That is the window I described earlier, and she climbed through it at precisely the moment it was open — while the growth was still there to fund the work, and before the cracks made the work impossible. By her own account the team was "already collapsing on all these spreadsheets," and rather than let the collapse become the operating model, she made it the reason to rebuild.
None of this required Home Depot to be Home Depot. It didn't come down to scale, or budget, or some structural advantage the rest of the market doesn't have. It came down to asking the right question early, looking honestly at the data, and being willing to change direction mid-build rather than defend the half-built house all the way to the ceiling. That is a choice available to very nearly everyone reading this — which is exactly why so few of us who make the opposite one can honestly call it inevitable.
Finishing the house.
The Wall Street Journal's Megan Graham recently wrote a fantastic article calling retail media a "gangly teenager" of an industry. The person who said it — Brian Monahan of Albertsons Media Collective — was admirably honest about what that means:
That last line is worth sitting with, because everything I've described in this piece — the deferred foundation, the manual workflows, the fragmented systems, the one-way doors — is exactly what the buyer feels as messy, inconsistent, and hard to work with. The half-built house was never only an internal problem we quietly absorb. It is the experience we hand to every advertiser who tries to spend with us. Growing up, in Monahan's terms, has very little to do with getting bigger. It is about becoming easier to do business with — and you cannot do that from inside a house you never finished building.
Some operators are already treating architecture as the work. Costco Wholesale's Mark Williamson stood up at an industry event earlier this year and walked a room full of competitors through his entire technology stack — not as a flex, but because being legible to the advertisers you want to attract is itself part of being easier to do business with. You can't be transparent about an architecture you're quietly embarrassed by, and you can't be legible when the honest answer to "how does this actually work?" is a spreadsheet, or three people doing it by hand in the background.
The room to do it is narrowing, too. WPP Media expects the industry to go through a "rationalization," in which only the largest and most technologically sophisticated networks win out, and the pressure on players without differentiated data or inventory becomes — in their word — "existential." That is the doom loop described at the scale of the entire market: growth slowing, advantage compounding for the few who built properly, and everyone else running out of the runway they'd need to catch up.
So if there is a place to start, it is the same place Stephenie started — and none of it requires being The Home Depot.
The retailers who grow up are the ones who decide, early and on purpose, to finish it.
Subscribe to the series. Share this if it resonated. And as always — comment, debate, push back. That's how we all get smarter.