Who should own retail media: the media agency, the commerce specialist or the performance shop? The honest answer in October 2026 is whoever understands the client's category best, because the channel no longer sits in one place. Retail media now runs through creator content, social shops and off-site audiences as much as through sponsored listings on a retailer's own site.
The evidence arrived this month. Digiday's Media Agency Report 2026 describes retail media networks as "becoming another key channel for creator partnerships", and notes that by June 2026 TikTok Shop was appearing in formal agency RFPs alongside Amazon and Walmart. When a social platform turns up in the same procurement document as two retailers, the old org chart stops working.
What has actually changed in retail media this year?
For most of the last decade, retail media was a bottom-of-funnel search channel. Someone in performance or ecommerce bought sponsored products, watched return on ad spend and reported it upwards. Brand and media teams barely looked.
Three shifts in the Digiday report break that picture:
- Creators have moved in. Digiday cites its first-quarter survey, in which 17% of respondents were partnering with influencers to create content for retail media. That made retail media the second most used channel for creator partnerships, after social.
- Data deals are formalising it. The report points to Omnicom Media's agreement with Meta to connect Walmart Connect purchase data with Instagram creator campaigns.
- Social commerce is being bought like retail. Go Fish Digital president David Dweck told Digiday his agency has been pushing a lot of clients into TikTok Shop, which he sees as "kind of the second biggest retail medium" outside Amazon.
Put those together and retail media is no longer one channel with one owner. It is a stretch of the funnel where search, social, creator and shopper marketing overlap, and every agency on the roster can make a plausible claim to it.
Which agency type should own retail media?
Each contender brings something real, and each has a blind spot. Here is how the claims stack up.
| Agency type | Its claim | Its blind spot | What it must prove |
|---|---|---|---|
| Media agency | Buying scale, audience planning, relationships with retailer networks | Treats retail media as inventory rather than a shopper problem | That the plan moves sales, not just reach |
| Commerce or marketplace specialist | Knows the digital shelf: listings, ratings, content, pricing, promotions | Can be too close to one retailer to see the whole category | That it can plan across retailers and social shops |
| Performance or search shop | Bid management, product feeds, measurement discipline | Optimises what is easy to count, which flatters branded search | Incremental sales, not retailer-reported returns |
| Creator or influencer agency | Content people actually watch, fluency in social commerce | Thinner on retail data and trade terms | That creator content converts at the shelf |
| In-house brand team | Owns the retailer relationship and the trade budget | Short on specialist talent and tooling | That it can run the channel without becoming an agency |
Notice what is missing from every row: the category view. Which competitors are spending where, what shoppers keep complaining about, which claims everyone is already making. Whoever brings that to the table tends to end up setting the strategy, whatever their badge says.
Why does nobody win retail media by standing still?
Michael Porter's Five Forces is an old business-school framework, but it explains this fight well. Profit in any market flows to whoever controls the scarce thing, and five pressures decide who that is: rivalry, new entrants, substitutes, buyer power and supplier power.
Apply it to retail media from the agency side of the desk:
- Supplier power is high. Retailers own the purchase data and the inventory. Agencies rent access.
- Rivalry is rising. Five types of team are now competing for the same budget line.
- Substitutes are coming. Self-serve tools and AI buying threaten the execution layer. Tinuiti's Harry Browne told Digiday that "agentic buying is a big topic, but it's a pretty ill-defined one right now". Ill-defined is not the same as harmless.
- Buyer power is firm. In Digiday's survey of 30 agency professionals, 36% said client media budgets were flat in 2026. Flat budgets mean any new channel is funded by taking money from someone else.
That is why stagnation is the silent killer of opportunity. The agency that keeps defining itself by the channel it buys will not lose retail media in one dramatic review. It will lose it a line item at a time, to whoever turned up with the sharper read of the category.
Where does agency value move when execution is automated?
The French trade title Stratégies made the point in its piece on a new era for media agencies. Sanja Vlaisavljevic, who runs the agency Lift, expects AI to gradually compress part of media execution and to shift agency value away from production itself. Her prescription, in our translation: understand the problem before looking for the media answer, and be able to challenge a brief, not just execute it.
Retail media is where that shift will bite first. Bid management, feed optimisation and reporting are exactly the tasks software is getting good at. What stays valuable is knowing why a competitor just doubled its creator content on one marketplace, or why shoppers in one region keep complaining about the same pack size.
How can an independent agency build a credible retail media position?
You do not need a holding-company data deal to compete. You need a view of the category that is fresher than anyone else's on the roster. A practical sequence:
- Map where the category actually sells. List every retailer, marketplace and social shop where the client and its competitors appear, including the ones the client ignores.
- Watch competitors continuously, not quarterly. Track their paid and organic content across those surfaces: claims, formats, creators, promotions and timing.
- Read the shopper conversation. Reviews, comments and creator threads show which tensions drive purchase. Turn them into search terms, product-page copy and creator briefs.
- Agree ownership in writing. Split strategy, execution and measurement across the roster explicitly. Ambiguity is how you end up doing the work while someone else takes the credit.
- Insist on incremental measurement. Retailer-reported returns are useful, but a network marking its own homework will always be generous. Push for holdout tests where you can.
- Report the market, not just the spend. A monthly note on what competitors did and what shoppers said does more for your seat at the table than another ROAS chart.
Steps two and three are where most teams stall, because doing them by hand takes days per client. That is the gap the Defy platform was built to close: Defy Monitor keeps live competitor, audience and conversation intelligence running on every client and category, so the category view exists before the retail media meeting rather than after it.
What should you tell a client who asks who should run their retail media?
Split the question in two. Who sets the strategy, and who runs the buying? They do not have to be the same agency.
Strategy belongs with whoever can explain the category: competitors, shoppers and the claims still up for grabs. Execution can sit with whoever has the best retailer access and tooling. Measurement should sit with someone who does not profit from the answer.
If your agency can only answer the second question, start building towards the first now. The roster will not wait for you.
