TL;DR
- IAB Europe released its third Attitudes to Retail Media Report on 3 November 2026, based on 150 cleaned respondents (91 buy-side, 59 sell-side) cut from 303 raw records, with Europe-only year-on-year trends.
- ROAS was chosen as a wanted metric by 49% of European buyers in 2026, down from 89% in 2025 (n=35, p<0.001); unified online and offline reporting (66%) now ranks first.
- 58% of buyers want multi-touch attribution but only 31% of sellers offer it, a gap the full report marks as significant (p=0.015).
- 68% of retail media sellers plan AI-powered campaign management within 12 months and 50% plan agentic commerce and conversational AI discovery tools.
IAB Europe released the third edition of its Attitudes to Retail Media Report on 3 November 2026, and its clearest finding is that European buyers have stopped treating return on ad spend (ROAS) as the measure that matters most. Only 49% of European buyers chose ROAS as a metric they want from retail media networks in 2026, down from 89% in 2025. Buyers now rank unified online and offline reporting (66%), new-to-brand or new-to-category measurement (63%) and incrementality (57%) above it. The survey also found that 58% of buyers want multi-touch attribution while only 31% of sellers offer it, and that 68% of sellers plan to build AI-powered campaign management within the next 12 months.
The figures below come from IAB Europe's 3 November announcement and from the 54-page report it links to. Where a number appears only in the full report, this article says so and gives the sample behind it, because several of the year-on-year comparisons rest on fewer than 50 respondents.
What IAB Europe published on 3 November 2026
IAB Europe, the European-level industry association for digital advertising, released the report from Brussels on 3 November 2026. It is the third annual wave of the study, following editions in 2024 and 2025, and the first to include what IAB Europe calls a meaningful sample from the Middle East and North Africa (MENA).
The report was developed by IAB Europe's Retail and Commerce Media Committee. Caras Consulting carried out the quantitative analysis, and Ascentinel conducted independent qualitative research, with LuckyCart and Mirakl Ads supporting the qualitative work as sponsors. Respondents included advertisers, agencies, retailers and retail media networks (RMNs).
Marie-Clare Puffett, Senior Director of Industry Development and Marketing at IAB Europe, said in the announcement that Retail Media is becoming firmly embedded in advertising strategies, but that there is "still work to do to build buyer confidence." She pointed to IAB Europe's Commerce Media Measurement Standards and its Retail Media Certification Programme as its response.
How the survey was run
The sample and method matter for reading the headline numbers, so they are set out here as the report states them.
- Format. An online survey distributed to IAB Europe members and the wider industry, with help from the network of National IABs and Federations.
- Cleaning. 303 raw records were reduced to 150 after removing screener drop-outs and four records that SurveyMonkey's Response Quality system flagged as poor quality.
- Split. 91 buy-side respondents (advertisers and agencies) and 59 sell-side respondents (retailers, media owners and RMNs), classified by company type and mutually exclusive.
- Regions. The analytical bases are Europe (110) and MENA (37). Eleven European respondents also hold MENA responsibility, so the two cannot be added together.
- Trends. Because the 2024 and 2025 waves were mostly European, every year-on-year comparison is made on European respondents only.
- Testing. Movements are called significant when a two-proportion z-test clears 95% confidence. Cuts with a base under 30 are suppressed or flagged as indicative, which is why the whole Europe versus MENA section is labelled as hypotheses.
- Interviews. Ascentinel interviewed 18 people, 11 on the buy side and 7 on the sell side.
Individual questions run on much smaller bases than 150. The ROAS comparison, for example, uses 46 European buyers in 2025 and 35 in 2026. The report also warns that some questions changed their option lists between waves, which can dilute each option's share.
ROAS loses its place at the top
The report calls the ROAS result the sharpest single-year movement in the dataset. Among European buyers, ROAS was chosen by 76% in 2024 and 89% in 2025, then 49% in 2026. The 41-point fall is significant at the 99.9% level (p<0.001), and the report says every other retained option in the same question moved within statistical noise.
The report flags one caveat itself: the measurement-metrics question added options in 2026, so part of the ROAS fall may reflect a longer list. It argues the move is still credible because ROAS alone collapsed while the options around it held. Across all buyers, including MENA, ROAS was chosen by 55% (n=40), against 49% on the European base.
The report's explanation is that ROAS is an efficiency ratio that counts sales that would often have happened anyway, and that buyers are moving to questions it cannot answer: whether a campaign brought in customers who are new to the brand, and whether it produced sales that would not otherwise have happened.
The measurement gap between buyers and sellers
The full report sets what buyers want against what sellers say they provide. The sell side still leads with delivery metrics. 76% of sellers offer campaign metrics such as reach and frequency and 73% offer ROAS, while only 43% of buyers want campaign metrics. Demand runs ahead of supply on unified online and offline reporting (68% of buyers against 40% of sellers) and on cross-product attribution (45% against 24%).
The report asks readers to treat those gaps as direction rather than exact size, because buyers ticked more options on average and the two lists differ. One gap is tested. On attribution models, 58% of buyers (n=36) say they ideally want multi-touch attribution, while 31% of sellers (n=42) say they offer it, a difference the report marks as significant at 95% (p=0.015). Single-touch is the model sellers offer most, at 45%, and 21% of sellers offer none of the listed models.
Sellers appear to know this. Enhancing the measurement offer is the sell side's most-cited development priority for the next 12 months at 63% (n=40), with AI capabilities and new ad products close behind at 58% each, and 45% intending to adopt industry measurement standards.
Barriers: measurement and data availability
61% of European buyers named a lack of measurement and reporting as a barrier to further investment, up from 48% in 2025 and now ranked first. The report treats that 13-point rise as an indicator rather than a significant change. The jump in concern about data availability, from 31% to 58% among European buyers, is significant (p=0.013). The report notes that the barrier list gained options in 2026, so it regards both rises as conservative.
On the all-buyer base of 41, lack of measurement leads at 61%, but lack of data, lack of standardisation and fragmentation of retail media networks follow at 59%, 59% and 56%. The report says that spread sits within the margin of error, and reads it as a tight group of four infrastructure barriers rather than one clear leader. Asked where standardisation is needed most, buyers named attribution measurement (79%) and media measurement (73%), well ahead of creative standards and ad formats (42%), on a base of 33.
Partnerships are longer, but maturity lags
71% of European buyers have partnered with retailers for more than a year, up from 51% in 2024 and 64% in 2025. The report calls this the clearest tested trend in the study, significant at 95% (p=0.039), on a 2026 base of 42. Yet buyers rate their own retail media maturity at 3.2 out of 5 on average, and only 12% place themselves at the top level, which the report calls "Sophisticated" (n=42).
Spending is still rising for most. 73% of buyers increased retail media investment over the past 12 months and 8% cut it (n=51). 63% of buyers work with four or more retail media networks and 18% with ten or more (n=49). When choosing a network, buyers put reporting and transparency (86%) and measurement options (70%) ahead of performance (68%), while CPMs sit sixth at 43% (n=44).
Tested year-on-year changes among European buyers
The report's appendix lists the movements it tested among European buyers. The table below shows the main ones.
| Measure (European buyers) | 2025 to 2026 | Test result |
|---|---|---|
| Want ROAS from networks | 89% to 49% | Significant (p<0.001) |
| Barrier: lack of data available | 31% to 58% | Significant (p=0.013) |
| Barrier: lack of measurement and reporting | 48% to 61% | Indicator, not significant |
| Opportunity: reach shoppers at point of sale | 82% to 62% | Significant (p=0.038) |
| Partnered with retailers for over a year | 64% to 71% (51% in 2024) | Significant 2024 to 2026 (p=0.039) |
AI plans on both sides
The AI questions were asked for the first time in 2026, so there is no trend. Among sellers (n=40), 68% plan to develop AI-powered campaign management and optimisation in the next 12 months, 50% plan AI-driven creative generation, and 50% plan capabilities for agentic commerce and conversational AI discovery. Only 13% are not looking to develop AI at all. The report says respondents were asked directly about product visibility inside large language models such as ChatGPT, Gemini, Copilot and Perplexity.
Buyers answered a single-choice version. Asked which one area will drive the biggest increase in investment, 39% chose AI-powered campaign management and 39% chose agentic commerce or conversational AI discovery, with AI creative generation at 22% (n=41). On the European base alone, agentic and conversational AI reads 33%.
The report is careful about what this shows. It says the study measures industry expectation and build plans, not consumer behaviour, and that it did not ask shoppers whether they use assistants to find products. It also notes that retail media measurement currently depends on watching a path through a retailer's own site, and that there is no agreed way yet to see, attribute or verify discovery that happens inside an assistant's answer.
What the report did not cover
- No Asian market data. The analytical bases are Europe and MENA. Some respondents hold responsibility for other regions, including APAC, but the report does not analyse those regions.
- No spend totals. It reports shares of respondents, not euro figures for the market.
- No network-level results. It does not rate or name individual retail media networks.
- No firm MENA conclusions. MENA bases mostly fall between 9 and 20 per question, and the report labels that section as hypotheses, not findings.
What this means for Thai marketers
The survey's analysed bases are Europe and MENA, so none of its numbers describe the Thai market. What transfers is the set of questions European buyers are now asking, and those questions apply to any retailer or marketplace ad platform a Thai brand buys from.
- Ask for more than ROAS. Before renewing a retail media budget, ask the network whether it can report new-to-brand customers and incremental sales, and whether it can join online and offline sales. A ROAS figure alone cannot show whether the ads produced sales that would not have happened anyway.
- Ask which attribution model is used. The report found single-touch is the model sellers offer most. Knowing whether results are single-touch or multi-touch changes how you compare one network with another.
- Check where the money comes from. The report found that half of buyers who re-allocate budget into retail media take it from brand advertising (n=42). If your own retail media budget comes from brand, the report's logic is that it should be judged on brand-level evidence rather than on last-click sales alone.
- Join retail media data with your own analytics. Networks report on their own estate. Matching their numbers with your site analytics and CRM is the practical way to test claims across several networks, and a working GA4 setup is the starting point for that.
For brands that sell through marketplaces and retailer sites, these checks sit alongside the wider ecommerce marketing plan rather than replacing it.
Frequently asked questions
When was the IAB Europe Attitudes to Retail Media Report 2026 released?
IAB Europe released it on 3 November 2026 in Brussels. It is the third annual edition and the first with a meaningful MENA sample.
How many people took part in the survey?
The report is based on 150 cleaned respondents, 91 buy-side and 59 sell-side, reduced from 303 raw records. Year-on-year comparisons use European respondents only, and many individual questions have bases of roughly 33 to 59.
Did buyers stop caring about ROAS?
Not entirely: 49% of European buyers still chose ROAS in 2026, down from 89% in 2025. It dropped out of the top four, behind unified online and offline reporting, new-to-brand measurement and incrementality.
Does the report include data on Thailand or Southeast Asia?
No. The analysed bases are Europe and MENA, so the figures should not be read as describing the Thai market. The questions buyers are asking can still be applied to Thai retail media buying.
What do retail media sellers plan to do with AI?
68% of sellers plan AI-powered campaign management and optimisation within 12 months, and 50% plan capabilities for agentic commerce and conversational AI discovery. The report says these are build plans, not measured changes in shopper behaviour.
The 2026 report shows European buyers asking retail media to prove incremental, cross-channel value, and sellers only partly ready to show it. If you want to review how your own marketplace and retail media results are measured, Relevant Audience can help you set up reporting that answers those questions.
