ChatGPT's Ad Go-To-Market is Backwards

From what I'm seeing in early testing, OpenAI is reaching out to large CPG brands with $60 CPMs (yeesh!) and a $200K minimum spend — no ROI framework, no measurement plan, no brand safety protocols, no competitive benchmarks.

Makes sense for revenue. But it’s backwards for proving the channel works and for what big CPG (the most conservative ad spenders) need to buy.

Here's why:

Large CPG brands have inconsistent organic AI signals. Heritage brands with dominant market share (bleach = Clorox; detergent = Tide) show up fine. But for anything nuanced — benefits-based queries, emerging categories, portfolio brands — AI defaults to challenger brands with clearer positioning.

Ask "what protein powder for gut health?" You'll get a DTC brand with specific claims over a Nestlé SKU buried in 2,000 products.

Paid placements can't fix weak organic signals. Just like Amazon: messy product page = ads won't save you.

And the timing couldn't be worse:

CPG is under pressure from every angle: rising input costs, slower consumption, retailer margin squeeze, tariffs, uncertain consumer spending. New channels mean more fragmented spend with lower ROI. Ad spending growth slowed to 4.6%.

I launched the first TikTok ad at Clorox. I remember those legal conversations. New channel, no precedent, unclear brand safety? Every dollar gets scrutinized. And that was with TikTok's measurement tools, at much lower entry costs, and during the golden age of CPG when budgets were healthy and growing.

Now? Budgets are tight and every investment needs immediate justification.

So OpenAI is stuck:

Big brands can afford $200K minimums but won't move without the infrastructure they need.

Challenger brands have stronger organic signals and would prove the channel works but can't afford the buy-in.

The brands that can afford it won't buy without proof. The brands that could prove it can't afford to test.

The smarter play? Start with challengers at lower pricing, prove ROI, build infrastructure, then scale to enterprise.

Instead, they'll spend months troubleshooting why expensive placements aren't driving results for brands AI doesn't understand.

This won't kill the channel but it definitely slows proof of concept.

Covering this + how challenger brands can actually win in AI visibility in this week's newsletter.

[Subscribe here → darlingstrategiesllc.com/newsletter]