Latest · AI

ChatGPT Ads Just Beat the Benchmark. Pay Attention.

October 10, 2026

Weight Watchers ran ChatGPT ads and came in 15.3% under CPA benchmark, according to DV. Rockerbox users can now log ChatGPT as a touchpoint in attribution, and there's a suitability pilot with OpenAI in the works. Translation: a new channel is showing early numbers nobody can fully audit yet, and the plumbing to measure it is being built in real time.

Here's why that matters for you. Every channel that's ever mattered in this business, paid search, Facebook, Google LSAs, was cheapest the week before everyone else noticed it worked. CPAs don't stay low because a channel is good. They stay low because it's empty.

So what do you actually do with one case study and no disclosed methodology. You don't move budget. You open a line item. Take a few thousand dollars that would've gone to a fourth ILS or a retargeting campaign nobody's checked in six months, and run a small, boring test on an AI-chat placement if your platform offers one. Track it like you'd track anything new: separate UTM, honest CPA math, no vibes.

One caveat worth saying plainly. "15.3% below benchmark" is DV's number, on DV's benchmark, from one advertiser's campaign. That's not nothing, but it's also not a category-wide trend yet. Early data from a measurement vendor with a stake in the pilot's success is a reason to test, not a reason to believe.

The upside of being early isn't that you're right. It's that you're cheap while you find out.

Source: PPC Land, October 7, 2026