Imagine that your skincare brand pays for an advert beside a face cream somebody was already looking for. They click. They buy. The dashboard credits your campaign. The spreadsheet smiles. The uncomfortable question remains: did you create that sale, or simply buy the credit for it?
Retail media, advertising bought through a retailer's website, app, store or customer-data network, is becoming harder for beauty and fashion marketers to ignore. On 4 November 2026, commerce-advertising company Criteo released results that neatly illustrated both the channel's momentum and the peril of confusing one number with another. Its retail-media advertiser spend grew, but revenue earned by Criteo from that business fell. Those are different measurements, not a contradiction and certainly not evidence that advertisers lost money (Criteo, 2026a; Criteo, 2026b).
For a brand planning its Christmas push, this is the better question than 'Should we try retail media?': What exactly will the retailer count as success, and how will we know whether any of those sales were genuinely additional?
The numbers sound strange until you read the labels
In the three months to 30 September 2026, Criteo reported $1.2 billion in total media spend across its operations, up 9% year on year at constant currency. On its earnings call, it said retail-media media spend had increased 22% year on year and reported approximately 4,700 brands using the platform. But retail-media revenue in Criteo's published accounts fell 23% to about $52 million (Criteo, 2026a; Criteo, 2026b). Independent trade analysis put the retail-media spend total at approximately $560 million (Petto, 2026).
Here's what those labels mean. Media spend is money advertisers put through campaigns. Platform revenue is what an advertising technology supplier recognises for providing its services. Advertiser return is what a brand gains from its investment. They are related, but none is a substitute for another.
Criteo attributed part of the gap in its results to changed scope with two retail-media clients. Its retail-media contribution after traffic acquisition costs fell 23%, but management said that, excluding these scope changes, contribution from the underlying client base grew 10% (Criteo, 2026a). That adjusted growth figure is context, not a cancellation of the reported decline.
The company also named beauty retailers Notino and Adore Beauty among partners added to its wider retail network. The announcement demonstrates where commerce advertising is expanding; it does not establish how much any one beauty brand sold, what those partners spent or whether the ads were profitable (Criteo, 2026a).
None of this means retail media is a bad buy. It means a fast-growing media channel cannot be judged from a technology provider's earnings headline. Brands have to understand their own attribution and their own margins.
First, what counts as retail-media advertising?
The most familiar version is a sponsored listing for a serum, mascara or fragrance appearing on a retailer's own product-results page. You can also buy retailer display placements, homepage features, emails, in-store screens and some campaigns that use a retailer's audience data beyond its own website. The inventory and targeting available will vary by retailer and buying agreement.
It is appealing because it brings a message near a shopping moment. A shopper comparing moisturisers may be more likely to consider a relevant skincare product than someone scrolling past a generic ad. Retailer first-party data can offer insight into real customer behaviour. But proximity to a purchase also creates a measurement trap: the shopper could have completed the same purchase without the paid placement.
For a young cosmetics brand, retail media might help it earn visibility against established products. For an established fragrance brand, it may reinforce an existing preference and take paid credit for customers who were going to purchase anyway. Those are hypotheses to test, not conclusions that can be inferred from a reported return-on-ad-spend figure.
The uncomfortable problem with 'attributed sales'
Suppose a retailer reports that an advert was followed by £10,000 in purchases. That's a useful observation, but it doesn't automatically prove the campaign caused £10,000 of new sales. How many purchases would have happened without the advert? Did the shopper search your brand by name? Was the item already in their basket? Did another channel generate the demand?
The UK's own advertiser trade body has documented the reporting challenge. ISBA and mediasense published their second Digital Retail Media Study on 8 September 2026, examining more than 200 capability points across five large UK retail-media networks. Their finding was not that retailers are incapable of measuring. Rather, different definitions, attribution methods, reporting services and access arrangements still obstruct like-for-like comparison (ISBA, 2026).
One dashboard might attribute a sale to an ad clicked within a set number of days. Another might include impressions without clicks, or use a different reporting window. A sale figure can be gross or net of cancellations and returns. Without those definitions, two attractive-looking reports might be answering entirely different questions.
The industry's response is underway. IAB Europe's Commerce Media Measurement Standards V2 set out consistent definitions for metrics including gross and net sales, new-to-brand reporting and incrementality, with a 30-day default lookback and flexibility for other windows (IAB Europe, 2026). That is useful progress, but a published standard is not a guarantee that every retailer dashboard already meets it.
Why this matters particularly to beauty and fashion
Beauty advertising makes a persuasive case visually: texture, shade, packaging, ritual, trust. Retail media can place those signals near a product choice, but it can also overvalue people who already love the brand. A cult cleanser with high repeat purchase may produce apparently excellent attributed sales because existing customers continue replenishing. The campaign may deserve some credit, but the amount needs evidence.
For fashion, returns complicate the picture. A promoted outfit can be sold today and returned next week. A dashboard that counts gross orders but does not deduct the cost of refunds, shipping, retailer fees or promotional discounts is not showing contribution profit. High order value does not necessarily mean high commercial value.
Retailer media also interacts with everything else you do. A beautiful paid social campaign could create desire, Google could help with research, a creator could build confidence and a sponsored retailer listing might receive the last click. The team buying that last placement should not be the only team defining success.
That is why a premium brand needs joined-up planning between creative strategy, marketing and technology , not isolated reports that compete to claim the same customer.
Seven questions to ask before buying the next placement
1. What is the exact audience? Existing buyers, category shoppers, new-to-brand prospects or a mixture? The answer changes the campaign objective. Acquisition and retention should not use identical success thresholds.
2. Which placements are included? Sponsored products in a results page are different from homepage display, emails and off-site media. Ask for placement-level costs and reporting instead of one blended number.
3. What creates an 'attributed sale'? Request the click and view attribution windows, the sale types included, and whether customer exposure is deduplicated across campaigns.
4. Are those sales gross or net? Ask about discounts, VAT treatment, cancellations, returns and shipping. Agree the retailer's definitions with your finance team before a campaign begins.
5. How much is genuinely incremental? For meaningful budgets, discuss a controlled experiment: for example, comparable groups where one group is intentionally not exposed to the campaign. A valid control needs care over audience overlap, seasonality and selection bias.
6. Can you see new-to-brand customers and category effects? Winning an order from a competing brand is different from moving an existing customer from your own full-price item to a discounted one. Each has different commercial implications.
7. Who owns the data and the learning? Will the retailer give you accessible data, campaign-level definitions and usable creative insights? If reporting disappears when the campaign ends, you are renting certainty as well as inventory.
The difference between reporting and proving
The IAB and IAB Europe published incremental-measurement guidance in November 2025 covering experiments, modelled counterfactuals, econometric methods and more pragmatic hybrid approaches. The important principle is not technical sophistication for its own sake. It is finding a credible estimate of the sales that would have occurred without the advertising (IAB, 2025).
For a small brand, that might begin with carefully selected product, audience or geographic comparisons, alongside disciplined pre/post analysis. Those simple comparisons can be misleading if a competitor launches, a holiday begins or stock runs out at the same time. For larger budgets, proper randomisation and independent analysis may make sense. There is no magical one-size-fits-all ROAS figure.
Consider a deliberately illustrative example, not a reported campaign result . A brand spends £2,000 and sees £10,000 of attributed sales. That suggests a five-times attributed ROAS. But if a defensible test indicates only £3,000 of those sales were additional, the incremental revenue-to-ad-spend ratio is 1.5 times. Subtract product costs, retailer fees, returns and fulfilment and the commercial answer may change again.
A dashboard that tells you precisely which purchases happened after an ad can still be imprecise about which purchases happened because of that ad. The distinction is not pedantry. It is the difference between buying growth and paying to collect a compliment.
A more beautiful media brief starts with a harder question
We would ask a beauty founder to start with three truths: the product truth that makes the ad worth noticing, the shopper need that makes the placement relevant and the measurement plan that will show whether the investment created value. The creative work should still be distinctive, desirable and platform-aware. Measurement is there to make good taste commercially useful, not to replace it with a spreadsheet.
Prioritise testing one placement and one clear shopping problem rather than buying every retailer format at once. Agree the margin model, capture a baseline, build clear products and landing pages, and record what you learn before the next spend decision. A tighter experiment can teach you more than a larger campaign with flattering but ambiguous results.
Our analysis of beauty discounting looks at the other half of profitable growth. For brands considering a retail partner or marketplace, see what Zalando's expansion means for control of the checkout . To bring creative direction, paid media and reporting together, explore The Digital Dollhouse's services or tell us about your next launch .
What should UK brands remember?
Is retail media worth buying for a small beauty brand? Potentially, if the retailer's shoppers match your customers, the placement addresses a real discovery problem and you can afford to test it with clear objectives. A platform is not inherently a good or bad investment.
Does a high ROAS prove a campaign created new sales? No. Attributed sales describe a relationship between exposure and purchase under a platform's rules. Incremental sales need a credible estimate of the no-ad alternative.
Did Criteo's retail-media revenue decline mean brands were losing money? No. Its revenue, advertisers' media spend and their own outcomes are different metrics. Criteo attributed part of its reported decline to changes in two client contracts.
References
Criteo (2026a) Criteo reports third quarter 2026 results , 4 November. Available at: Criteo Investor Relations (Accessed: 5 November 2026).
Criteo (2026b) Q3 2026 prepared remarks , 4 November. Available at: MarketScreener, reproduction of company remarks (Accessed: 5 November 2026).
ISBA (2026) ISBA and mediasense report establishes first capability baseline for UK digital retail media networks , 8 September. Available at: ISBA research announcement (Accessed: 5 November 2026).
IAB Europe (2026) IAB Europe releases updated commerce (incl. retail) media standards and flexi ad sizes guidelines , January. Available at: IAB Europe (Accessed: 5 November 2026).
IAB (2025) Guidelines for Incremental Measurement in Commerce Media , 3 November. Available at: Interactive Advertising Bureau (Accessed: 5 November 2026).
Petto, A. (2026) 'Criteo retail media spend rises as revenue falls', Grocery Trade News , 5 November. Available at: Grocery Trade News (Accessed: 5 November 2026).