Stat of the Week. US commerce media advertising revenue reached $63.4 billion in 2025, up 18.0% year over year, though the growth rate is down from 23.0% the year before. (IAB and PwC, April 2026.)
Many marketing budgets still treat commerce media as just a retailer negotiation, debated with a few major chains and placed somewhere between trade spend and paid media. This view has been outdated for two years, and by 2025, sticking with it started to have real costs. Now, the category includes more sellers outside of the retail sector, shifting the focus for marketing leaders from simply sizing up the opportunity to choosing the right suppliers. The key question is which transaction-data partners deserve a real commitment and on what terms, especially as many of your current partners are quietly becoming media sellers. The best approach is to work with fewer partners, invest more deeply in them, and agree on reporting terms before making any commitments.
Commerce media has outgrown the retail sector
It's important to look at the scale before discussing strategy. The Interactive Advertising Bureau (IAB) and PwC report that US commerce media will reach $63.4 billion in 2025, up $9.7 billion from the previous year. The entire US internet advertising market hit $294.6 billion, growing by 13.9% (IAB and PwC, April 2026). When a channel grows faster than the overall internet advertising market for two years in a row, it is no longer just an experiment.
The same report notes that growth is slowing. Commerce media is expected to grow by 18.0%, down from 23.0% the previous year. IAB says this is partly because it's harder to maintain high growth on a larger base, and partly because brands are asking retailers to prove that the spending is truly incremental. Another IAB forecast, based on a January survey of 205 US ad decision-makers, predicts commerce media will grow 12.1% in 2026, compared to 9.5% for all US advertising, which includes TV, radio, print, and direct mail as well as digital (IAB, January 2026). Growing about a quarter faster than such a broad market, on a $63 billion base, shows this is now a solid part of the plan.
The real change is in how the category is defined. IAB now says commerce media is advertising powered by commerce data across on-site, off-site, and in-store environments, plus new and growing players beyond retail (IAB and PwC, April 2026). The key point is that the definition now focuses on who holds the data, not just who owns a store.
Two assets make a business a media seller
If you look at what the new entrants have in common, the list is short. To qualify, a business needs just two things.
First, they need a record of what people actually bought. Purchase history is a clear fact. A card issuer, airline, or delivery marketplace can target buyers more confidently than a publisher who guesses based on what someone reads. This also helps settle any disputes after a campaign, since the seller can see if a purchase happened.
Second, there must be a moment when the customer is making a decision. This could be at checkout, on a booking screen, in a statement, or during a service call—times when the customer is focused and can see their options. This makes the advertising space more valuable than the same impression elsewhere.
Almost any business that requires a login and a payment step meets these criteria, and the market is responding. In June, AdExchanger reported that The Trade Desk created a travel and hospitality media network with Uber Advertising, Booking.com, United’s Kinective Media, and Marriott Media (AdExchanger, June 2026). One goal is to give advertisers access to travel data they might not have considered before, such as marketing inside hotels. For example, supermarkets can't offer insights about when someone is planning a trip.
A few years ago, buying this kind of media would have been much harder. In 2025, programmatic advertising revenue reached $162.4 billion, up 20.5% (IAB and PwC, April 2026). This automation allows media teams to add a mid-size network without needing more staff.
Breadth is the expensive part
The main challenge in this channel is no longer supply. IAB points out that what slows brands down is the work involved in running campaigns across different retailer systems, separate buying platforms, and fragmented workflows, as well as inconsistent measurement (IAB and PwC, April 2026). There are also questions about how well incrementality is tested. Each of these costs is paid for every partner, but media plans often overlook this.
A packaged-food company spent a year testing nine networks: three grocery chains, a delivery marketplace, a convenience retailer, a card-linked offer platform, and three regional accounts. This made sense, since a learning budget should be spread out while there is still something to learn. Each network reported a positive return by its own standards, so the year looked successful. The problem appeared in the next planning cycle, when the team had to decide which network to scale up. Looking back, the four smallest networks took up about half of the analysts' time but only accounted for 6% of the spend, since each required its own creative specs, portal, and manual product mapping. These four networks didn't underperform—they just never received enough funding to justify the effort. The next year, the same budget went to two networks, each getting four times the spend, with reporting terms agreed in advance. This finally gave the team a clear answer they could use.
The buyer on the other side may be AI-agents
There is an important question in this category: will the customer still be the one making the choices? Accenture’s 2026 survey of 25,590 people in 16 countries found that 74% would let an AI agent handle routine tasks, like negotiating a deal or resolving a complaint, as long as the agent follows their instructions (Accenture, 2026). This shows people are willing to delegate errands, but fewer are ready to give up decision-making. Thirty-two percent would let an agent choose what to buy if they still make the payment, and only 9% are open to fully autonomous purchases. Willingness drops sharply when it comes to spending money.
There are two possible paths, but both lead to the same approach. If agent-mediated buying stays close to the current 9%, commerce networks will keep selling directly to people, and nothing changes. If more people let agents make buying decisions, then businesses with purchase records and payment steps will become the main channels for agents, turning today’s media partners into tomorrow’s customer routes. In either case, spreading your budget thinly across many sellers does not pay off. It is better to know two partners well enough to negotiate effectively.
Fund fewer partners, and set the terms first
You can address most of these issues before the next planning cycle starts, and you do not need a new budget line to do it.
Begin by listing which of your current partners have already become media sellers, even if you have not officially changed the relationship. The bank behind your co-brand card, the marketplace delivering your product, and the platform your customers use to book travel are all now part of this category. Each of these agreements was originally made for a different purpose. This list is often longer than your media plan shows, and it is where your negotiating power already exists.
Next, decide what a network needs to report for its results to be compared with everything else you buy. Agreeing on these definitions before signing is very different from asking for them later, and it is one of the few points of leverage a mid-size advertiser has with a large seller.
This category is still new enough that its terms are being shaped, and those terms will be set by advertisers who commit deeply to a few partners. Which non-retail network has surprised you this year? Reply in the comments with your answer.
Sources
IAB and PwC, “Internet Advertising Revenue Report, Full-Year 2025,” April 2026 (US market; all figures realized. Commerce media $63.4B, +18.0%, an increase of $9.7B, with the report stating the rate is “down from 23.0% in the prior year”; total internet advertising $294.6B, +13.9%; programmatic $162.4B, +20.5%. The definition quoted, “on-site, off-site, and in-store environments, as well as new, growing entrants beyond retail,” and the operational-constraint passage are the report’s own words).
IAB, “2026 Outlook Study: A Snapshot of U.S. Ad Spend, Opportunities, and Strategies for Growth,” January 2026 (Projection, not a realized result: forecast US ad-spend growth of 9.5% for 2026 — a total-advertising basis that includes linear TV, radio, print, OOH, and direct mail — with commerce media at 12.1%. Email survey of US buy-side ad investment decision-makers, n=205 (44% brands, 53% agencies, 3% other), fielded 11/24/25–1/16/26; the by-channel commerce media estimate is answered by n=115 of them. IAB states the 9.5% is lifted by cyclical events and that removing them lowers growth to a 7.1%–7.8% range, so the article’s “roughly a quarter faster” is the conservative reading. Cited to the study PDF rather than the January 28 press release, per the content plan’s follow-links-to-the-primary rule; the two agree on every figure. This is also the study the revenue report’s separate “30% faster than the overall market” line is footnoted to, so the forecast is cited here rather than to the revenue report).
AdExchanger, “The Trade Desk Forms A Travel And Hospitality Media Network,” June 17, 2026 (trade reporting by James Hercher; names Uber Advertising, Booking.com, United Airline’s Kinective Media and MARRIOTT MEDIA as partners. No figures are given in the article, and none are cited from it).
Accenture, “Talk to my AI agent: The new rules of brand value,” Consumer Pulse 2026 (n=25,590 consumers across 16 countries). Stated openness, not observed behavior: 74% would delegate routine tasks “provided the agent acts strictly on instruction,” 32% “would let an agent decide what to buy, provided that they make the payment themselves,” and 9% “are open to an AI agent making fully autonomous purchases.” The 74% appears on the same page in a second, different framing (trusting an agent more than a best friend) — the delegation framing is the one used here, and the body says in-line what this 74% measures.