The following is a guest piece written by AppsFlyer President and General Manager Brian Quinn. Opinions are the author’s own. 

ChatGPT ads mark the beginning of large-scale ad monetization for large language models. It’s hardly a stretch to say that it’s going to work in the short term. The audience is significant and growing, it is rich with intent signals, and the strategic rationale for marketers to test this surface is clear. Early reports are coming in from brands testing it across many industry verticals. As with any fast-moving sector of our industry, people don’t want to be left out.

The real question for marketers is not whether ChatGPT ads will attract early spending, but whether they will become a durable part of the performance media mix.

History suggests that the answer will depend less on audience and engagement and more on measurement architecture.

The inevitable maturity curve

Every new channel follows a familiar arc. It launches with scale and momentum. Marketers test, and budgets flow. For a period of time, it can grow on that momentum alone.

Eventually, the novelty wears off, and the standards rise. The key question becomes how the channel performs relative to the rest of the plan, and whether it delivers incremental outcomes that justify reallocating dollars from other channels.

That is the point at which a promising surface either matures into a dependable budget line or stalls as a perpetual test.

Mobile was one of the earliest large-scale proving grounds for this dynamic. As usage accelerated, performance investment followed. At the same time, mobile introduced signal fragmentation, platform IDs and walled environments that made traditional web-based measurement insufficient. The market adapted by building independent attribution and incrementality frameworks that reconcile performance across platforms. Third-party measurement became a prerequisite for serious performance budgets.

Today, connected TV illustrates a later stage of the same pattern. Linear was a brand medium, and CTV has largely inherited that role, with stronger performance capabilities as audiences moved to streaming. But while viewers shifted quickly, budgets have moved a bit more slowly, in part because measurement remained fragmented across platforms and devices. Mobile was different; it supplanted the web, which was already a performance channel, so it had to prove itself accordingly. 

LLMs are likely to face the same standard as they begin to capture activity that once belonged to search. Historically, search has been one of the clearest performance environments in marketing. That means LLM ads will ultimately be judged the same way: whether they can be measured, compared, and justified against the rest of the mix.

That gap matters to advertisers because it shows that consumer adoption alone does not guarantee budget maturity. In other words, the hurdle for new channels is not demand but comparability. ChatGPT ads are now entering this same cycle.

Even closed ecosystems open for comparability

The most successful advertising platforms have learned this lesson. Meta, TikTok, Google, Snap, etc., are frequently described as walled gardens, a characterization that remains accurate in important ways. At the same time, none of them operate in complete isolation. Each maintains select partner programs with independent measurement providers. 

These integrations are structured and controlled, and notable precisely because they are the exception to the rule. Walled gardens did not integrate independent measurement out of a preference for openness. They did so because mature channels require cross-channel comparability. Performance budgets move when marketers can evaluate outcomes using consistent frameworks, not when the platforms grade the homework.

For media buyers, planners and measurement leads, that comparability is what makes budget shifts defensible inside the organization. Audience scale and engagement built these ecosystems. Strategic, third-party measurement integrations helped normalize them within the broader omnichannel mix and made them easier to evaluate alongside other channels using the same decision-making framework.

LLMs have the opportunity to do the same.

A chance to build it correctly from the start

LLMs are not simply another ad surface. They represent a new mode of user behavior. Conversational interaction compresses discovery, evaluation and action into a single interface. In doing so, it has the potential to divert traffic from traditional search, brand websites and other established channels. That makes LLM advertising inherently full funnel and inherently disruptive.

When a channel affects not only where ads run but also how consumers navigate the broader ecosystem, integration becomes more important. As answer engines redirect intent that once flowed through search and web, advertisers will need to understand how those shifts change performance across their entire portfolio. They will need to see LLM exposure and downstream behavior in the same view as search, social, apps and CTV. Without that consistent picture, it will be difficult to determine whether LLM advertising is expanding the pie or redistributing existing demand.

That is why the industry should treat measurement infrastructure not as a secondary consideration, but as a condition of long-term channel growth.

ChatGPT and other LLM platforms have the advantage of building monetization in an era where these expectations are already established. If LLM advertising is integrated with independent measurement partners from the outset and supports standardized cross-channel analysis, it can mature without the friction that has slowed other channels. 

Dollars will flow in the short term on momentum alone. But for agencies, brands and platforms alike, the long-term opportunity depends on whether LLM ads can be evaluated with the same rigor as the rest of the media plan. If that happens, LLM advertising will no longer be an experimental line item. It will become a normalized part of how modern media budgets are built.

If it does not, the market may still be large, but it will be harder for the channel to become truly endemic.