OpenAI Ads: My Honest Review After Running Them Since Beta

A real OpenAI ad rendered inside ChatGPT, exactly the placement I've been running since June.

Why I jumped on OpenAI ads early

I've been running OpenAI ads since June 5, back when the beta first opened up. It's a placement ad, and I've been pointing it at people who'd be interested in my marketing services. Being early on a new ad platform is usually where the cheapest, least competitive clicks live, so I wanted to see for myself.

Here's my honest read after weeks of live data, not a press release.

They behave like a display ad, not a search ad

My first impression is that ChatGPT ads act a lot more like Facebook Display or Google Display ads than like a search campaign. You get a lot of impressions and comparatively few clicks. The clicks you do get, though, seem to perform pretty well.

My real numbers so far: 2,950 impressions against 27 clicks, a 0.92 percent click-through rate. That's roughly in line with what I see on Meta, not some breakout number, but the quality of the clicks has been solid so far.

Conversion tracking is still rough

The other thing I noticed is that setting up conversion actions in OpenAI's ads manager is more cumbersome than it should be. With Google and Facebook, you're usually adding one or two scripts and defining a couple of conversion actions, and both platforms give you a tool that shows what a conversion event looks like firing in a live browser. OpenAI doesn't have that tool yet.

The pixel itself feels early, which makes sense for a beta. I'd expect this to get easier fast, since Google and Meta both solved the same problem years ago.

Cheap for now, average once it's not

Click-through rates on OpenAI ads are about average. I don't see a meaningful difference from what I get on Facebook ads.

The real advantage right now is price. Because the platform is new and inventory isn't crowded yet, you don't have to pay as much for a click as you would on a more mature platform, and that math won't hold forever as more advertisers pile in.

The AI targeting promise doesn't hold up yet

Here's my actual criticism, and it's the part that matters most. OpenAI's whole pitch is that AI-native advertising means better targeting: ads that show up in the exact moment someone is thinking through a decision, informed by richer context than a keyword ever gave you. I have not seen that play out.

My click-through rate is average. My results are not meaningfully better than what I get running the exact same offer on Facebook or Google. If the promise is "AI understands intent, so your ads perform better," the data I have doesn't back that up, at least not yet.

Right now OpenAI ads look like a normal ad platform wearing an AI story, not a fundamentally better one. Cheap inventory is a real, temporary advantage. Smarter targeting is not something I can confirm from where I'm sitting.

Who should actually try this

My honest recommendation: for most small businesses, start with Facebook, and start with Google once you're more established. That's still where the volume and the targeting maturity are.

Where OpenAI ads earn a spot is with clients who are willing to experiment with around a thousand dollars a month. Think of it as a budget version of Facebook ads, worth testing while it's cheap, not something to bet your whole budget on yet.

Where this fits in a real strategy

Paid ads, OpenAI included, are the short-term lever in a marketing plan: fast to turn on, sensitive to the platform's own demand market, and done the moment you stop paying. I cover the full platform mix, including OpenAI ads, as part of my paid ads management services.

If you want the play-by-play on the other platforms, I've also written up my honest take on Facebook and Instagram ads and on where Google Ads budgets actually leak.