The growth story is settled. Retail media has moved from an interesting experiment to a permanent budget line, and the direction of spend is no longer in question — we mapped that flow in where retail media budgets are moving in 2026. What's far less settled is whether the advertisers funding that growth can actually tell it's working. Industry research this year found roughly three-quarters of retail media advertisers naming incrementality as their single biggest measurement challenge — while only around 15% described themselves as very or extremely effective at measuring it. That gap, not the growth curve, is the real story for anyone spending the money.
Why the money is going there (the fair case)
Start with why this is happening, because the case is genuinely strong and worth stating properly before picking it apart.
Retailers hold something that became dramatically more valuable as third-party tracking eroded: verified first-party purchase data. That enables targeting most publishers can't match and, crucially, closed-loop reporting — the ability to connect an ad impression to an actual product sale rather than an inferred one. In a market where attribution keeps getting harder, that's a compelling pitch. Add reaching shoppers at the moment of purchase intent, and the fact that retail media carries far better margins for retailers than selling groceries, and you have a channel both sides had strong reasons to build quickly.
It also does more than convert. Research this year suggests retail media campaigns generate measurable lifts in brand image and purchase intent relative to other media, which undercuts the lazy framing of it as a pure bottom-funnel tactic. The case is real. The problems sit alongside it, not instead of it.
Problem one: the measurement gap
Here's the uncomfortable pairing that should shape how you approach this channel.
The gap Around 75% of advertisers say incrementality is their biggest measurement challenge. About 15% feel able to measure it. And roughly seven in ten report hitting their goals — against numbers they don't trust.
That last clause is the part worth sitting with. Satisfaction is high, confidence in measurement is low, and both statements come from the same survey population. It's a pattern that should be familiar to anyone who has watched a channel scale on self-reported success — the network grades its own homework, the numbers look good, and the budget grows.
Surveys in 2026 also found a majority of US advertisers citing lack of measurement standardisation as their biggest retail media problem, and Forrester research has put the share of commerce media decision-makers treating attribution improvement as a high or critical priority in the high eighties. This isn't a fringe complaint. It's the defining operational issue of the channel.
Landingi
Drag-and-drop landing page builder with reusable Smart Sections, A/B testing and 400+ templates.
Best for: Marketers and agencies building pixel-perfect landing pages without code.
Problem two: fragmentation is structural, not temporary
The second issue is the one most likely to consume your team's actual hours. Advertisers now work across an average of around six retail media networks, with projections suggesting that number could roughly double by the end of 2026. Each network brings its own platform, login, ad formats, naming conventions, reporting, and support relationship.
No other digital channel looks like this. Paid search is effectively one dominant platform; social advertising is a handful. Retail media is dozens to hundreds of networks globally, and the industry consensus has shifted from treating that as a growing pain to accepting it as a permanent structural feature. The practical consequence is that operational cost scales badly — doubling your network count roughly doubles the manual work without doubling the insight.
Problem three: you can't compare across networks
This is where fragmentation becomes a measurement failure rather than an inconvenience. Networks measure on their own terms: one may use a seven-day view-through window, another same-day click-only. Amazon reports by product, other platforms by audience or keyword. Event definitions, identifiers and taxonomies differ.
Which means a ROAS figure from one network is not arithmetically comparable to a ROAS figure from another. You can put both numbers in a spreadsheet, but the comparison is meaningless — and the budget reallocation decision you'd base on it is unsupported. It's also why feeding retail media data into unified modelling is so difficult: the inputs don't share a language.
| What you get | What you need |
|---|---|
| Attributed sales, self-reported | Incremental sales — what wouldn't have happened otherwise |
| Network-specific ROAS | A comparable figure across networks |
| Last-click, short-window attribution | Contribution across the whole journey |
| Performance inside one walled garden | Cross-channel effect, including offline |
| Whether the campaign hit its target | Whether the budget belonged there at all |
Problem four: the incrementality question
The structural issue underneath all of it deserves saying plainly. Retail media reaches shoppers at or near the point of purchase — someone already on the retailer's site, already searching for a product category, often already intending to buy something in it.
That's the source of both its high reported returns and its measurement problem. A meaningful share of attributed sales would likely have occurred anyway, and the closed-loop reporting that makes retail media attractive is also what makes it flattering: every sale that follows an impression gets counted, including the ones you'd have won regardless. Readers of our piece on retargeting done right will recognise the pattern exactly — advertising to people already leaning toward purchase produces impressive-looking numbers by construction.
None of this means retail media doesn't work. It means the reported figure and the incremental figure are different numbers, and only one of them should drive budget decisions.
Problem five: whose budget is this?
A quieter issue that causes real internal friction. Retail media has historically sat between trade spend — money negotiated with a retailer as a cost of doing business — and marketing budget, which is expected to demonstrate return. Plenty of programmes are funded from the first and judged by the standards of the second.
Industry bodies published framework guidance during 2026 aimed at separating genuinely measurable retail media from legacy trade spend, which should help clarify things over time. But the internal question still needs an explicit answer in your own organisation, because it determines what "good" looks like. If it's trade spend, the benchmark is retailer relationship and shelf presence. If it's media, it has to compete with every other channel on return — a comparison that gets sharper in a year when budgets are effectively flat and every increase is funded by a cut somewhere else.
What good advertisers are doing
1. Define your own metrics. If networks won't standardise, impose internal consistency — agree what an impression, a new-to-brand sale and an incremental unit mean to you, and normalise everything into that.
2. Run your own incrementality tests. Geo experiments and audience holdouts you control, rather than relying on network-reported attribution.
3. Feed it into modelling. Marketing mix modelling handles walled-garden channels better than click-based attribution because it works on aggregates.
4. Consolidate networks deliberately. More networks means more operational cost; concentrate where you have genuine scale rather than being present everywhere.
5. Agree the budget question internally. Trade or media — decide, then apply that standard consistently.
Notice that four of the five are things you do, not things you ask the networks for.
That last observation matters. Waiting for the industry to standardise has been a losing strategy for several years running, and the honest assessment from analysts is that full standardisation is unlikely in the near term. The advertisers handling this well built their own measurement layer instead — the same self-reliance argued for in the return of marketing mix modelling and in getting conversion data you can actually trust.
The one asymmetry to keep in view
Worth naming plainly, because it rarely appears in vendor-authored coverage: when you buy retail media, you're funding the advertising business of a company that also competes with you for the end customer's attention, and supplying data about your own performance in the process.
That isn't a reason to avoid the channel — the reach and purchase-point intent are real, and the retailer relationship usually matters commercially anyway. It's a reason to negotiate consciously, to avoid becoming dependent on any single network, and to keep building demand you own outright alongside it. Product presence and conversion quality on those platforms remain yours to control, which is why the fundamentals in product page optimisation carry over directly to retail media performance.
The question to actually ask
If there's one change to make in how you evaluate this channel, it's replacing the question. Stop asking "what's our retail media ROAS?" — a number your networks will happily supply, calculated their way, and which won't compare to anything else you're running.
Ask instead: what would have happened without it? That question is harder, requires testing you have to run yourself, and often produces a less flattering answer. It's also the only version that survives a serious budget review. For a fuller picture of how these networks work and why brands are drawn to them in the first place, our overview of why retail media networks are growing so fast covers the mechanics — and getting the commerce side of the operation right is where an experienced ecommerce marketing team earns its place.
The bottom line
Retail media's surge is real, the underlying proposition is sound, and it belongs in most commerce-driven media plans. But the channel has scaled faster than the industry's ability to evaluate it, which puts advertisers in an awkward position: increasing investment in the medium they can least confidently assess. Take the opportunity, but take it with your eyes open — standardise your own metrics because nobody else will, run incrementality tests you control rather than accepting attributed sales, consolidate networks instead of collecting them, and settle internally whether this is trade money or media money. The advertisers who'll look smart in two years aren't the ones who spent the most on retail media. They're the ones who could prove what it did.
Spending on retail media without knowing what it's returning?
We standardise measurement across networks, run the incrementality tests, and show you what your commerce media is genuinely worth.
Explore Performance Marketing →Frequently asked questions
Why is retail media growing so fast?
Three things converged. Retailers hold first-party purchase data that became far more valuable as third-party tracking eroded, giving them targeting and closed-loop sales reporting most publishers can't match. Advertisers get to reach shoppers at the moment of purchase intent, which converts well. And retail media is highly profitable for retailers compared with selling goods, so they've invested heavily in building networks. The result is a channel that grew quickly because it suited both sides — not necessarily because its returns were rigorously proven.
What is the biggest challenge with retail media in 2026?
Measurement, and specifically incrementality. Industry survey work in 2026 found around three-quarters of advertisers naming incrementality as their biggest measurement challenge, while only about 15% described themselves as very or extremely effective at measuring retail media performance. Roughly the same proportion said they strongly trust their measurement. The channel has scaled faster than the ability to evaluate it, which means many advertisers are increasing investment in the medium they can least confidently assess.
Why can't you compare ROAS across retail media networks?
Because networks measure differently and report on their own terms. One may use a seven-day view-through window while another uses same-day click-only, and each defines events, identifiers and taxonomies its own way. A ROAS figure from one network therefore isn't arithmetically comparable to another's. Industry surveys have found a majority of US advertisers citing the lack of measurement standardisation as their biggest retail media challenge, and it's the main reason budget reallocation across networks is so difficult to justify.
Is retail media advertising incremental?
Often less than reported figures suggest, and it's genuinely difficult to know. Retail media reaches shoppers at the point of purchase, so a meaningful share of attributed sales would likely have happened anyway — the same structural flattery that affects retargeting. Distinguishing incremental sales from baseline sales requires testing the networks generally don't provide: geo experiments, audience holdouts, and modelling run independently. Some networks have begun launching incrementality measurement, but advertiser-controlled testing remains the more credible route.
Is retail media trade spend or marketing budget?
Historically it has sat awkwardly between the two, often funded from trade budgets negotiated with retailers while being judged against marketing performance standards. That ambiguity causes real problems, because trade spend is typically a cost of doing business with a retailer while media spend is expected to demonstrate return. Industry bodies published framework guidance in 2026 aimed at separating measurable retail media from legacy trade spend, which should help — but the internal question of which budget owns it still needs answering explicitly.
