OpenAI’s measurement gaps are keeping ChatGPT ads budgets at test level

Seven months after OpenAI began selling ads in ChatGPT, marketers are still grappling with how to measure their effectiveness. The issue has become a major point of discussion in conversations about OpenAI's performance as an ad seller. Part of the frustration stems from the perception that OpenAI hasn't developed many advanced measurement tools.

Marketers' own wariness about trusting the company with their data further complicates the matter. Agencies running campaigns note significant gaps in OpenAI's ad reporting capabilities, which are not as common on Google or Meta platforms. Accuracast conducted a U.K. lead-gen campaign exclusively on ChatGPT and observed no conversions two weeks later, while Google and Meta platforms typically report more conversions than they actually deliver.

Accuracast's CEO Farhad Divecha mentioned that Google and Meta platforms often overstate their conversions. Similarly, Accuracast's platform only recognized around 20 conversions, while it showed 100 clicks. OpenAI's support team at the time couldn't provide an explanation for the discrepancy. Such issues are not universal, though.

Executives from AdRoll, Jellyfish, and another independent U.S. ad agency reported that they have not encountered click discrepancies on OpenAI's platform. For Web Guide Partner, the timing of insights is also a concern; conversions registered on ChatGPT arrive 24 to 36 hours later compared to a few hours on Google. The reporting on ChatGPT fails to include demographic or prompt data.