Aimerce Blog

Is OpenAI Ads Manager Actually Working for DTC Brands?

16 September 2026

Is OpenAI Ads Manager Actually Working for DTC Brands?

Quick Answer: Early data on OpenAI Ads looks different than a typical ad channel: click-through rate is low by search standards, pricing is still underpriced relative to where it will land, and a meaningful share of consumers say chatbot advertising makes them trust a brand less. None of that means it isn't working. It means the usual metrics don't apply yet.

Key Takeaways

OpenAI dropped the $50,000 minimum ad spend and opened self-serve buying, and early CPCs reflect a temporary pricing inefficiency, not the long-term cost of the channel.
A 0.91% click-through rate looks weak against search, but the ad appears after ChatGPT's answer rather than instead of one, so a non-click following a complete answer isn't automatically a failure.
Categories where people already research conversationally before buying, software, education, professional services, considered DTC purchases, fit naturally. Restricted and impulse categories don't.
Attribution is still maturing. A LiveRamp Conversions API Hub partnership, announced June 10, 2026, extends measurement beyond browser tracking, but building your own interim attribution, post-purchase surveys, dedicated landing pages, geo holdouts, is still the more reliable approach today.
Consumer trust in chatbot advertising is measurably lower than trust in more established ad formats right now, a real cost of testing early that's smallest where a product already feels native to the conversation.

Is OpenAI Ads Manager actually working right now?

It's working well enough to be worth testing, but not in a way that maps cleanly onto how search or social ads get judged, which is exactly why it looks confusing from the outside.

OpenAI opened self-serve ad buying on ChatGPT roughly two months ago, dropping the $50,000 minimum spend that had gated the earlier beta. Aimerce built a server-side Conversions API integration the moment beta access opened, which we've covered in detail separately, so this piece isn't about setup mechanics. It's about the more interesting question: given what early data actually shows, should a DTC brand be testing this channel right now, and if so, how should it judge whether the test is working.

Should you test OpenAI Ads while it's still underpriced?

Yes, if your category fits, because early CPCs reflect a temporary inefficiency in an auction that hasn't found its real price yet, not the channel's long-term cost.

Self-serve buying killed the old $50,000 minimum, which means the auction is still thin relative to where it'll land once more advertisers pile in. That's the same pattern every new ad auction goes through, early Meta and early TikTok both had a window where CPCs were cheap simply because demand hadn't caught up to supply yet. That window closes as more advertisers show up, and it tends to close faster than people expect once a channel proves it converts.

Which product categories actually fit OpenAI Ads?

The categories where people already research conversationally before buying: software, education, professional services, and considered DTC purchases. Impulse and restricted categories don't fit the same way.

Which Categories Fit OpenAI Ads Right Now?

Category type	Fit	Why
Software / SaaS	Strong	High-consideration purchase path matches how people already use ChatGPT to research options
Education / courses	Strong	Same research-heavy, comparison-driven decision pattern
Professional services	Strong	Trust and research-driven, longer consideration window
Considered DTC purchases (higher-ticket, comparison-shopped goods)	Strong	Shoppers already research conversationally before committing
Impulse-purchase DTC (low-cost, trend-driven, discovery-based)	Weak	Conversational research doesn't match how these purchases actually happen
Restricted categories under OpenAI's ad policy	Skip	Policy risk regardless of fit

Why is OpenAI Ads' click-through rate so low, and does that mean it's underperforming?

Not necessarily, because the ad appears after ChatGPT has already delivered a complete answer, not in place of one, which makes a low click rate structurally different from a low click rate on a search ad.

A 0.91% CTR looks weak next to search benchmarks, where the ad often is the answer a user is looking for. Here, the user already got their answer before the ad ever appeared. A non-click in that context can just as easily mean the answer already did its job, not that the ad failed to earn attention. Judging this channel by CTR alone imports an assumption from search that doesn't hold in a conversational context.

What should you measure instead of CTR?

Dismissal rate, ad-unit engagement, and cost per qualified visit, since those measure whether the ad unit itself is doing its job rather than whether it interrupted something successfully.

Dismissal rate tells you whether people are actively rejecting the ad unit, which is a much stronger negative signal than simply not clicking. Ad-unit engagement captures interaction that doesn't necessarily end in a click but still indicates the ad registered. Cost per qualified visit ties spend back to something that actually matters downstream, a visit with real purchase intent, rather than a raw click that may or may not represent one.

How should you track conversions until OpenAI's own attribution matures?

Build your own interim measurement, post-purchase surveys, dedicated landing pages, and geo holdouts, rather than waiting for click-based attribution to fully mature.

OpenAI's measurement infrastructure is improving quickly. LiveRamp's Conversions API Hub, announced as a partnership on June 10, 2026, extends measurement beyond browser-based tracking by matching conversion events server-to-server [1][2]. But coverage of the rollout indicates it's still maturing, with early emphasis on offline and in-store purchase matching rather than full parity with established channels. Until that's further along, a dedicated landing page lets you isolate traffic cleanly, a post-purchase survey question ("how did you hear about us") catches attribution that click-based tracking would otherwise miss, and a geo holdout gives you an actual causal read on incremental impact rather than a modeled guess.

What does it mean to match your ad to the conversation?

It means investing in context signals so the ad reads as a relevant suggestion inside the conversation, not an interruption dropped into it regardless of what was being discussed.

An ad that ignores the context of the conversation reads the way a bad retargeting ad reads on any other platform, present, but clearly not paying attention to what the person actually asked. The brands that get this right treat the ad placement as an extension of the answer rather than a paid insert competing with it.

What is the "trust discount," and how big a risk is it really?

It's the brand-trust cost of appearing in an ad format consumers haven't fully accepted yet, and it varies significantly depending on how natural your product already feels in a conversational context.

Multiple recent surveys point in the same direction: a large share of consumers report lower trust in AI chatbot advertising specifically, higher than the skepticism typically seen toward ads on more established formats. That's a real cost to price in when deciding whether to test early, not just a footnote. The size of that cost isn't uniform, though. It's smallest for products people would naturally ask a chatbot about anyway, and largest for brands whose presence in the conversation itself feels like the interruption.

Should your brand test OpenAI Ads now? A quick decision framework

If you sell in a research-heavy, considered-purchase category: worth testing now, while CPCs still reflect an underpriced auction rather than the channel's real long-term cost.

If you sell impulse or trend-driven products: the channel's underlying behavior doesn't match how your buyers actually shop. Watching from the sidelines a bit longer costs you less than testing does.

If your product already feels native to a conversational context: the trust discount is smaller for you specifically than the survey averages suggest.

If your brand is trust-sensitive or still relatively unfamiliar to the buyer: treat the trust discount as a real cost of testing early, not a rounding error.

Common mistakes to avoid

Judging performance by CTR alone. The ad's placement, after a complete answer rather than instead of one, makes a lower CTR structurally different from a search ad's CTR.
Waiting for OpenAI's own attribution to fully mature before tracking anything. Post-purchase surveys, dedicated landing pages, and geo holdouts can measure impact today, without waiting on the platform's own tooling.
Testing a conversational ad format in a category where people don't research conversationally in the first place. Fit matters more than budget here.
Ignoring the trust cost entirely. Early presence in an unproven ad format carries real brand risk, and it's worth sizing that risk against your specific product and audience rather than assuming it's negligible.

FAQ

Is OpenAI Ads Manager worth testing yet? For research-heavy, considered-purchase categories, yes, especially while CPCs still reflect an underpriced early auction. For impulse or trend-driven categories, the underlying conversational research behavior doesn't match how those purchases happen, making the channel a weaker fit right now.

Why is OpenAI Ads' click-through rate so much lower than search advertising? Because the ad appears after ChatGPT has already delivered a complete answer, not in place of one. A user who doesn't click may already have what they needed from the answer itself, which makes CTR a less reliable success signal here than it is for search ads.

What's the best way to measure OpenAI Ads performance right now? Dismissal rate, ad-unit engagement, and cost per qualified visit, paired with independent measurement like post-purchase surveys, dedicated landing pages, or geo holdouts, rather than relying on CTR or platform-reported clicks alone.

Which product categories should be cautious about testing OpenAI Ads? Impulse-purchase and trend-driven categories, since the format is built around conversational research, a behavior pattern that doesn't match how those purchases typically happen. Restricted categories under OpenAI's ad policy should skip it regardless of fit.

What is LiveRamp's role in OpenAI Ads measurement? LiveRamp and OpenAI announced a Conversions API Hub partnership on June 10, 2026, letting advertisers match conversion events server-to-server rather than relying solely on browser-based tracking. Coverage of the rollout indicates it currently emphasizes offline and in-store purchase matching, with broader online conversion parity still developing.

Does running OpenAI Ads carry a brand trust risk? Yes, measurably. By one figure, 57% of consumers report lower trust in chatbot advertising specifically compared to more established ad formats. That risk isn't uniform, it's smaller for products that already feel native to a conversational context and larger for brands whose ad presence itself feels like an interruption.

Does running OpenAI Ads carry a brand trust risk? Multiple recent surveys indicate consumers report meaningfully lower trust in chatbot advertising specifically compared to more established ad formats. That risk isn't uniform, it's smaller for products that already feel native to a conversational context and larger for brands whose ad presence itself feels like an interruption.

Why are OpenAI Ads' CPCs so cheap right now? Because self-serve buying only recently removed the $50,000 minimum spend that had limited the advertiser pool, leaving an auction that's thinner than it will be once more advertisers enter. Early pricing reflects that temporary imbalance, not the channel's likely long-term cost.