MarketingThis week in marketing

OpenAI's measurement gap is capping ChatGPT ad budgets

Seven months after ads went live inside ChatGPT, agencies say conversions go missing, arrive late or fail to match their own tracking, and brands are keeping spend at test level. For marketing leaders, this is a live decision about how much money moves into a channel that still grades its own homework.

A red chat bubble balloon held low on a short rope while a surveyor measures it with a theodolite.

Listen to the podcast

8 min

Chapters

Key takeaways

  • Define your KPIs and an independent measurement framework in writing before any ChatGPT ad spend, rather than accepting platform reporting at face value.
  • Run a geographic holdout, switching the channel off in matched regions, so results do not depend on OpenAI's own conversion counting.
  • Put OpenAI's conversion tracking terms in front of your legal team now with a decision date, since you cannot argue about data you refuse to collect.
  • Keep ChatGPT ads funded from test money until OpenAI names third-party measurement partners with a timeline, not just an intention.
  • Track whether conversion-optimised bidding produces a stable cost per acquisition in your own category before shifting budget from Google or Meta.

Episode page and full transcript →

According to Digiday, seven months after OpenAI started selling ads in ChatGPT, marketers are still asking how they are supposed to measure them, and it has become the sticking point in almost every conversation the publication has had about the company as an ad seller, driven partly by a sense that OpenAI has not built much beyond the basics and partly by marketers' reluctance to hand over their data.

The specifics are uncomfortable. Digiday reports that the agency Accuracast ran a UK lead generation campaign on ChatGPT alone and watched form fills arrive while the platform still showed zero conversions two weeks later, and that the same agency saw around 100 clicks reported where its own tracking picked up 20, with OpenAI's support team unable to explain why. AdExchanger picked up the click discrepancy the next day, noting OpenAI could not explain it. Where conversions do register, Web Guide Partner told Digiday they land 24 to 36 hours late against a few hours on Google, and reporting carries no demographic or prompt data.

What is contested: the experience is not universal. Digiday says execs at AdRoll, Jellyfish and another independent US agency had not seen click discrepancies. The problems also only apply to advertisers who have switched conversion tracking on, and many have not, because doing so means signing OpenAI's tracking terms that plenty of legal teams refuse on liability grounds. That leaves a loop: advertisers want Google-grade optimisation, but OpenAI cannot deliver it without the conversion data those same advertisers will not share.

Why it matters for budget owners: cost per click now accounts for most ChatGPT ad spend, according to Digiday, and Jellyfish's Victor Batista said the bigger shift comes when OpenAI optimises toward cost per acquisition or return on ad spend, which requires OpenAI to trust its own numbers first; until then advertisers bid with little to go on.

What to watch: independent measurement. OpenAI's ads and monetisation lead Asad Awan told a press briefing that third-party measurement was in the works without partners or a timeline. LiveRamp has since become the first independent ad tech company piping conversion data into OpenAI's conversions API, scoped to the US and select mutual clients. AppsFlyer and Adjust are signed as mobile measurement partners.

Finished reading?

Validate your read to earn XP and feed your radar.