Criteo Becomes ChatGPT's First Ad Tech Partner — But the Measurement Gap Looms Large

Criteo announced on March 2, 2026 that it is the first advertising technology partner to integrate with OpenAI's advertising pilot in ChatGPT, bringing performance-driven commerce media into conversational AI for the first time. The integration allows brands to buy ChatGPT ad placements through Criteo's commerce media platform — a significant milestone for both the ad tech company and OpenAI's nascent advertising business.

But the announcement also throws into sharp relief the measurement challenges that make ChatGPT one of the most expensive and least measurable ad channels available to marketers today.

How We Got Here

OpenAI confirmed its advertising plans on January 16, 2026, after months of industry speculation and an internal "code red" directive in December 2025 that had temporarily paused advertising development. By February 6, the company was accepting advertisers, with early participants including Target, Ford, Mrs. Meyer's, and Adobe. Holding companies WPP, Omnicom, and Dentsu signed on as launch partners. Ads first appeared in ChatGPT on February 9 for logged-in adult users on the Free and Go subscription tiers.

The ads are clearly labeled as sponsored and visually separated from ChatGPT's organic responses. OpenAI matches ads to the topic of the conversation, past chats, and prior ad interactions — so a user researching recipes might see ads for meal kits or grocery delivery. Crucially, OpenAI states that ads do not influence ChatGPT's answers, which are optimized solely for helpfulness.

What Criteo Brings to the Table

Criteo's integration is more than just a distribution deal. The company's commerce media platform sits on data from more than $1 trillion in annual commerce sales, giving it a unique ability to connect ad exposure to downstream purchase behavior — at least on the platforms where it has tracking in place.

As Digiday reported, Criteo found that users referred to retail websites from LLM platforms like ChatGPT convert at approximately 1.5x the rate of other referral channels, based on a sample of 500 US retailers observed in February 2026. That figure suggests users arriving from conversational AI are further along in purchase consideration than typical referral audiences — they've already researched, compared, and are closer to buying.

The partnership also ties into Criteo's broader push into what it calls "agentic commerce." In early February, the company introduced an Agentic Commerce Recommendation Service that demonstrated up to 60% improvement in recommendation relevancy compared to approaches relying solely on product descriptions, using a Model Context Protocol implementation that lets AI assistants access Criteo's recommendation infrastructure.

The Measurement Problem

Here's where things get difficult for measurement professionals. ChatGPT ads currently offer only high-level reporting: total impressions and clicks. There is no conversion tracking, no attribution pixel, no conversion API, and no UTM integration. Advertisers are paying a $60 CPM with a $200,000 minimum commitment and receiving impression counts without downstream conversion data.

For context, that $60 CPM is roughly three times Meta's average rates, comparable to premium CTV inventory from Netflix and Hulu ($40-65 CPM), and significantly above Google Display Network ads (around $3 CPM). As Search Engine Land reported, the premium pricing comes with far less measurement infrastructure than any of those mature platforms offer.

The attribution gap is structural, not just a beta limitation. In a chat-based environment where research, comparison, and decision-making can all happen without ever clicking an ad, traditional click-based attribution breaks down. A user might ask ChatGPT about the best running shoes, see an ad for Nike, and then buy from Nike.com three days later after a branded search — but there's currently no way to connect those dots.

What Measurement Teams Should Do

For brands considering ChatGPT ads, the measurement playbook looks more like early-stage CTV or podcast advertising than performance display. Several approaches can fill the attribution gap:

The irony is that while ChatGPT ads lack the real-time, click-level attribution that digital marketers are accustomed to, the recommended measurement approaches — incrementality testing, MMM, and brand lift studies — are exactly the methodologies that the broader industry is shifting toward as cookie-based attribution declines. ChatGPT may be forcing advertisers to adopt the measurement future ahead of schedule.

The Financial Stakes

The financial motivation behind OpenAI's advertising push is significant. Despite reaching a $20 billion annualized revenue run rate, OpenAI projects losses of approximately $14 billion in 2026, with cumulative losses of $115 billion through 2029. Evercore ISI analyst Mark Mahaney projects that ChatGPT advertising could become a $25 billion annual business by 2030. Internal OpenAI documents forecast $1 billion in revenue from "free user monetization" in 2026 alone.

Those projections depend on advertisers being willing to spend at premium CPMs with limited measurement — or on OpenAI building out the measurement stack fast enough to justify the investment. Criteo's partnership is a step toward the latter, but there's a long road between commerce media integration and the kind of closed-loop attribution that performance marketers demand.

What Comes Next

The Criteo deal signals that OpenAI is serious about building programmatic infrastructure, not just running a managed insertion-order business. But for measurement professionals, the key question isn't whether ChatGPT ads work — Criteo's 1.5x conversion data suggests they might — it's whether advertisers can prove they work within the ROI frameworks their organizations require.

Until OpenAI ships a conversion API, implements proper UTM support, or opens up to third-party measurement partners, ChatGPT advertising will remain what one industry observer called "the most expensive brand awareness play in digital." For the brands with the budget and the measurement sophistication to run proper incrementality tests, that may be a bet worth making. For everyone else, the measurement gap makes it hard to justify the $200,000 entry price.