ChatGPT Ads: why we're not spending there yet.
OpenAI is rolling out ads in ChatGPT. Should you advertise there? Our honest answer: not yet — and here's the math behind it.
Whenever a new ad channel appears, the same question follows: "should we get in?" For ChatGPT Ads, our answer right now is: no — not yet. Not on principle, but because the numbers don't support it yet.
The promise
A huge, highly engaged user base. Conversations with genuine buying intent. And the pull of novelty — first-mover fantasy. It sounds tempting.
Why the product is immature
A performance channel is more than reach. Meta and Google spent 15 years building what you don't see:
- No mature conversion optimisation — the algorithm doesn't reliably learn who buys yet.
- No robust tracking / no mature pixel-and-CAPI chain, no clean attribution.
- Little inventory and format control, unproven creative formats, untested CPMs.
A channel without conversion optimisation is a billboard with a login.
Why the math doesn't work yet
Without conversion optimisation and attribution, you're buying reach on faith. For a performance budget that's a clear no: you pay the early-adopter tax without the early-mover advantage, because the channel can't find your buyers yet. The same budget returns more, predictably, on Meta or TikTok today.
When we'll change our mind
We're watching three signals: a real conversion API, real optimisation toward purchases, and reliable attribution. The moment case studies with repeatable CAC show up, we'll test small and controlled — like we do with any new channel. Until then, your money stays where it works.
Want this on your account?
30 minutes, we'll show you exactly how we'd run it for your brand. No pitch deck.
Book a Growth Audit →