What is retail media, and when is it worth building?

Retail media is almost always explained from the brand's side: how do I buy visibility inside my retail partner's shop, and what does it cost. The more interesting question sits on the other side of the table. A retailer already owns reach, first-party data and advertising surfaces, and today mostly gives them away. What does it take to turn that into a revenue stream, and when is it not worth it? This piece answers that from the operator's chair.

01What is retail media when you are the retailer?

Retail media is the monetisation of three things a retailer already owns: its reach, its first-party data, and its advertising surfaces. Seen from the operator’s side rather than the brand’s, it is a revenue stream built on assets that most retailers currently give away.

Reach means returning visitors in the shop, the app, the newsletter, the store. First-party data means what was bought, viewed and returned, tied to an account or loyalty ID. Surfaces means search results, category pages, the homepage, product detail pages, screens at the point of sale, leaflets and circulars.

The money comes from supplier marketing budgets. Much of it was already flowing under different names: co-op funds, listing fees, promotional contributions, negotiated annually and almost never measured. Retail media does not add that money to the system. It makes it auctionable and provable. That is exactly why suppliers go along with it, and why buying departments sometimes resist.

02Why is retail building this now?

Retail is building retail media now because of margin, not because of the trend. Advertising revenue on owned surfaces carries margins far above the retail core business, and that gap is what makes a CFO listen.

Boston Consulting Group put the contrast in numbers in 2021: advertising revenue on owned surfaces can reach margins as high as 80 percent, against a typical in-store margin of 10 to 20 percent. Its 2022 follow-up, produced with Google and based on interviews with more than 50 retail media leaders, breaks that down: onsite inventory at 70 to 90 percent margin, offsite at 20 to 40 percent after media and agency costs. Both are estimates from 2021 and 2022, and they should be read as orders of magnitude rather than planning figures. The order of magnitude is enough. One additional euro of ad revenue contributes several times what an additional euro of merchandise revenue contributes.

Source: IAB Europe, AdEx Benchmark 2025

The European market points the same way. In its AdEx Benchmark 2025, published in July 2026, IAB Europe reports €13.3 billion in retail media, up 16.7 percent, inside a digital market of €131.1 billion growing 10.5 percent. Retail media passed 10 percent of European digital ad spend for the first time. A separate, more broadly scoped IAB Europe release from October 2025 puts 2024 at €13.7 billion with 21.1 percent growth and forecasts €28.8 billion by 2028. The two series are not interchangeable because the measurement scope differs. If you argue with these numbers internally, say which one you mean.

There is a second driver that gets less airtime. The less reliable third-party tracking becomes, the more valuable logged-in, volunteered customer data gets. In IAB Europe’s 2025 buyer survey, 85 percent of buyers name access to retailer first-party data as a key reason to invest. A retailer with an account system sits on an asset no ad network can rebuild.

03What are you actually selling?

The inventory splits into three groups that behave completely differently in the P&L: onsite, offsite and in-store. This is also what a retail media network looks like in practice, a stack of these three inventory classes wired to one auction and one report.

Onsite is your own digital surface area: paid placements in search and category, display on home and category pages, brand pages, recommendation trays, newsletter modules. This is where the margin lives, where measurement is cleanest, and where everyone starts. In IAB Europe’s 2025 survey, more than 90 percent of buyers put at least 41 percent of their retail media budget into it.

Offsite is bought-in reach, targeted with your own audience segments: social, video, programmatic display, search ads. Its share is climbing fast. The proportion of buyers allocating at least 41 percent of budget to offsite jumped from 30 to 46 percent in the same survey, and BCG estimated in 2022 that offsite was growing 35 percent a year against 22 percent for onsite. Offsite also dilutes margin, because the inventory has to be purchased.

In-store covers screens, shelf space, displays, receipts, trolleys, leaflets and circulars. It gets discussed constantly and, as of 2025, still attracts the least budget, because measurement there is the hardest.

Our recommendation is uncomfortably clear: if you are starting, start with paid search placements. Not because they are the most attractive ad format, but because they are the only one whose effect can be proven without an argument. Everything further up the funnel sells more easily once you can put numbers on the table from the bottom of it.

04What do you need in place before the first campaign runs?

You need seven building blocks before the first retail media campaign runs, and none of them is optional. They run from clean product data through to a named team, in that order, because each depends on the one before it.

A supplier who learns their campaign ran while the advertised product was largely out of stock will negotiate differently next time. That is precisely why the in-stock number belongs in the report, and why measurement is a building block rather than a nice-to-have.

05Where do retail media programmes fail?

Retail media programmes fail on four recurring patterns: the conflict between buying and media sales, ad load, missing standards, and minimum budgets the supplier base cannot support. Each one is organisational before it is technical.

The conflict between buying and media sales. Buying negotiates terms, the media unit sells visibility, and the supplier faces two people from the same company both asking for money for shelf presence. Settle this before launch or you will sell the same thing twice and lose the supplier on the third attempt.

Ad load. Every surface sold is a surface no longer working for relevance. A search results page where the first four positions are paid sells more advertising this quarter and less merchandise next year. You need a binding ceiling that does not depend on the media unit’s revenue target.

Missing standards. In IAB Europe’s 2025 survey, 53 percent of respondents named a lack of standardisation and 51 percent named network fragmentation as the biggest brakes on growth. For a single retailer that translates directly: invent your own metric definitions and your numbers become incomparable to a brand’s media planner, and incomparable numbers lose to comparable ones.

Minimum budgets your supplier base cannot support. If you have 30 relevant suppliers and set entry budgets only five of them can afford, you do not have a business model. You have five clients and a concentration risk.

06What does retail media measure, and what does it not?

Return on ad spend is the revenue generated for every advertising dollar, a ratio, not a cause. Retail media measures two layers: standard ad metrics, and the retail-specific numbers that carry the real value. What it does not measure, in your own dashboard, is incrementality.

The standard metrics are quickly listed: impressions, reach, frequency, clicks, click-through rate, cost per thousand, cost per click. Then the retail-specific ones that carry the actual value: product views, sell-through, average selling price, in-stock rate.

What the buy side wants is unambiguous. In IAB Europe’s 2025 survey, return on ad spend is the most in-demand metric at 88 percent, and when buyers choose a retail media partner, transparency leads at 82 percent, ahead of performance at 76 percent and measurement options at 75 percent.

It measures purchases that followed an ad, on a surface where the customer was already shopping. The cleanest evidence for this is not a market study but a field experiment in Econometrica: Blake, Nosko and Tadelis switched paid search ads off experimentally at a large online marketplace and measured the difference. The causal effect turned out to be a fraction of what non-experimental analysis showed. For ads on the firm’s own brand name there was no measurable short-term effect at all, and across frequent buyers the return was negative.

Our position: build control groups in from the start, geographic or audience-based. It costs reach and looks like a disadvantage in year one. It is the difference between a business that survives a large supplier’s first serious audit and one that permanently loses the budget in it.

07How far along is the German market?

The German market is further along than the debate suggests, and narrower in shape. Adoption of offsite media is broad, but investment concentrates near the bottom of the funnel, and proof is the constraint rather than inventory.

EHI Retail Institute, together with Google, surveyed specialists from eleven retailers and eight brands in spring 2024. It is a qualitative study with a small sample, so percentages on it are indications rather than market figures. As indications they still hold up: 91 percent of the retailers surveyed were already using offsite media, with just under 40 percent of ad spend going to search engine marketing, followed by social, newsletters and apps at roughly 20 percent each.

The more interesting finding in the same study is not about channels but about the gap. Brands invest mostly in lower-funnel, sales-driving activity and barely in the upper funnel, and half of the brands surveyed criticised the absence of standards for metrics and measurability. Translated into an instruction for the operator: the bottleneck is not inventory, it is proof. A retailer who can demonstrate effect transparently does not only sell more, they sell more expensive inventory.

08Does this only apply to consumer retail?

No. Retail media applies to any business with reach, identity and demand data, and B2B operators often meet those preconditions better than a consumer shop does.

A technical distributor, an industrial marketplace, a spare parts platform or a B2B specialist dealer often qualifies more cleanly than a consumer retailer. Reach is smaller, but identity is complete because everything runs through accounts, and order value is orders of magnitude higher. Anonymous traffic, the biggest measurement problem in B2C, barely exists there.

We work in both worlds, in fashion retail and with manufacturers of self-adhesive materials and labelling equipment, and the gap is smaller than expected. A distributor listing 400 suppliers, of which 40 actively compete for visibility in the same product category, has exactly the auction dynamic a retail media business runs on. What is usually missing is not the preconditions. It is the question.

09How do you know whether you have a retail media business at all?

You have a retail media business when reach, supplier competition, clean data, joined-up measurement and a named owner all hold at once. Answer these five questions honestly and the gaps will be obvious.

If three of those five are a no, build the foundation first. A retail media programme on messy master data with no measurement chain produces exactly one outcome: disappointed suppliers who then stay away for years.

10Where to start

If you want to know whether your reach and your data can carry an advertising business of their own, let’s talk for 30 minutes. We will run the five questions above against your numbers and tell you plainly whether building is worth it, or whether the foundation comes first.