The Retail Media Exodus Has Started, And It’s Not Slowing Down
Twelve retailers migrated from Google Ad Manager to Topsort in the past year, and that’s just one platform they’re moving from. That number is not a sales stat. It's a signal.
Something structural is changing in retail media, and migration is becoming the clearest evidence of it.
Retail Media Has Outgrown Its First Stack
This matters because the stakes have changed. US advertisers are projected to spend more than $71 billion on retail media in 2026, up from $60 billion in 2025. Retail media is no longer a side business or an experimental revenue line. It is a core P&L item, and it is being managed accordingly.
For years, many retailers launched media businesses on top of first-generation ad servers, managed-service platforms, and point solutions that were built to prove demand, not to compound it.
The goal was once activation, but that has now shifted. The next phase of retail media is not about getting live. That means:
- Compounding performance over time
- Expanding formats across surfaces
- Standardizing measurement
- Growing revenue without growing headcount at the same rate
The infrastructure requirements for that phase are fundamentally different from what most early setups were built to handle.
Topsort has been framing this shift plainly: retail media is moving from tools to infrastructure. The migration data suggests the market is arriving at the same conclusion.
Why This Shift Is Happening Now
The pressure behind that shift is now visible everywhere, and there’s three forces driving it.
Scale
Retail media now represents a meaningful share of advertiser budgets. As spend grows, systems need to handle more campaigns, more advertisers, and more complexity, without degrading. Once retail media becomes a real revenue line, infrastructure choices stop being feature comparisons. They become the same category of decision as payments, search, or core commerce systems. The standard changes.
Complexity
Running a retail media business today means more than serving ads. It means managing product catalogs and inventory signals, running auctions across placements, supporting advertiser self-serve workflows, and delivering consistent, trustworthy reporting. Many retailers are doing this across multiple stitched-together systems, which creates compounding operational overhead that gets harder to absorb as programs scale.
Revenue Pressure
Retail media is no longer allowed to be experimental. It is expected to grow predictably, improve performance quarter over quarter, and demonstrate clear value to advertisers. At that stage, infrastructure inefficiencies stop being invisible. They become the reason growth stalls.
That is what a maturing category looks like. Once the market starts standardizing outcomes, the conversation shifts from "what can launch" to "what can hold up."
The Core Misalignment with Legacy Systems
This is where older systems begin to show their limits, and it comes down to what they were originally designed to monetize.
The platforms that powered retail media's first wave share a common assumption: that the hard part is getting ads live. At scale, the hard part is something else entirely. Publisher systems were built to monetize attention: impressions, page slots, campaign delivery, direct sales workflows.That is the economic model they were optimized for.
Retail media monetizes something different: transactions, product relevance, inventory dynamics, conversion probability, seller competition. The underlying logic is not the same.
A system designed for content monetization can support retail media at a basic level. But as programs scale, the gap between what the system was built for and what the business now requires becomes harder to bridge. More integrations. More operational work. More manual intervention to get reporting clean enough to show advertisers. The workarounds accumulate.
Why This Is Migration, Not Optimization
The shift happening now is not about improving performance within the same setup. It is about replacing the underlying system.
Retailers making these changes are not adding features. They are consolidating systems, rebuilding campaign structures, and rethinking how their media businesses operate.
Migration at this level involves:
- Auditing inventory and placements
- Rebuilding campaign structures
- Mapping targeting and reporting logic
- Running parallel systems before full cutover
That level of effort only happens when the existing setup no longer fits the business.
What retailers are moving toward follows a clear pattern. They want infrastructure that sits closer to commerce: systems that understand product catalogs and search intent natively, run unified auctions across formats and surfaces, and give advertisers transparent, real-time reporting they can actually trust. They want to reduce operational complexity, not add to it. And they want self-serve tools that let advertisers move without requiring manual ad ops at every step.
In short: they are replacing systems built to serve ads with infrastructure expected to make economic decisions. That is a different product category, even when the surface looks similar.
What Retailers Are Moving Toward
As these changes happen, a pattern is emerging in what retailers are looking for.
In practice, that means moving toward systems that are easier to integrate, more aligned with product-driven advertising, and better suited to handle auctions, reporting, and advertiser workflows in one place. They are replacing stitched stacks with infrastructure that can sit closer to the monetization core.
What This Signals
The most important thing about twelve retailers migrating in a single year is not the number itself. It is what the decision represents.
Each migration is a conclusion: the old setup is no longer enough. Not for the scale the business has reached. Not for the advertiser expectations now in play. Not for the measurement standards the category is moving toward.
That is what migration signals. And in infrastructure categories, once that shift starts, it does not reverse.
