By
Austin Coker
ChatGPT Ads went from a private pilot with roughly $200,000 upfront commitments to a self-serve auction with a $500 starter credit in under a year. If you manage paid media budget, the question is no longer whether this inventory is real. It is whether your category is allowed in, what a click actually costs, and whether the people you sell to ever see an ad inside ChatGPT at all.
The short version: most consumer and ecommerce advertisers can open an account today and start on clicks, with OpenAI suggesting a $3 to $5 per click starting point. Most B2B and SaaS companies can’t buy their way in effectively yet, because the ad policy is written around consumer categories and because paid ChatGPT plans carry no ads. Below is the eligibility picture, the real cost structure, how targeting works, how the channel compares to Google and Meta, and what to do instead if your buyers use ad-free plans.
We also look at a theory spreading among advertisers right now: that ChatGPT Ads send you organic ChatGPT traffic too, and how much of that is real.
ChatGPT Ads are sponsored placements that appear below a ChatGPT response, labeled as ads and kept separate from the answer itself. Each ad includes the advertiser name, favicon, a title, ad copy, a landing page and an image, per OpenAI’s documentation. They’re shown to logged-in adults on the Free and Go plans, and OpenAI has stated that ads don’t influence the answer, that conversations are kept private from advertisers, and that user data isn’t sold. Plus, Pro, Business and Enterprise accounts see no ads at all.
Eligibility is the first filter, and it rules out more businesses than the pricing does.
OpenAI’s ad policies, updated September 10, 2026, state that ads are primarily limited to consumer categories such as household goods, local services, travel, and digital products or education, with room to expand later.
Prohibited categories include dating and adult content, alcohol, tobacco and vaping, gambling, cannabis, THC and psychedelics, political advertising, credit repair and debt settlement, unsupported health claims, and diet pills. The ad, the image and the landing page all have to comply, not just the creative. OpenAI also reserves the right to decline advertisers for business or competitive reasons.
Three categories are restricted and reviewed case by case in the US: financial services, health services and legal services. Licensing proof may be required.
There are also exclusions on the user side. Ads don’t run near health, mental health or political topics, and accounts belonging to under-18 users don’t see ads.
Here’s the practical verdict by business type as of October 2026:
Cost has moved fast, and the number you heard six months ago is probably wrong.
The early managed pilot reportedly ran around a $60 CPM with roughly $200,000 in upfront commitments, according to eMarketer. By April 2026 the minimum had been cut to $50,000 and select ads moved to click-based pricing.
Today there’s no minimum spend listed on the self-serve Ads Manager, and new advertisers receive $500 in ad credit after spending $500.
You’re billed by impression or by valid click. You set a max CPM bid for reach campaigns and a max CPC bid for click campaigns, and OpenAI recommends starting in the $3 to $5 per click range, per its help documentation. Prices are set in a relevance-weighted second-price auction, so creative and landing page quality affect what you actually pay.
For budgeting, a useful starting frame is a four- to six-week test at a spend level that produces at least a few hundred clicks at your bid ceiling. Anything smaller tells you about delivery, not performance.
OpenAI picks ads based on the context and intent of the conversation, your landing page, your ad text, and advertiser “context hints,” which are topics rather than exact-match keywords. OpenAI is explicit that hints don’t guarantee placement.
On top of that, the August 2026 update added product feeds, platform and geographic targeting, and custom audiences. As of October 2026, qualifying advertisers can also add negative phrases.
If you’re used to keyword-level control in Google Ads, plan for less of it. Your targeting levers are topic hints, feed quality and a landing page that makes the match to the user’s intent obvious. It helps to understand how one question expands into many related subtopics, which we covered in our guide to query fan-out.
Campaign objectives are Views, Clicks and Conversions. The standard placement sits below the response. A visual ad format shown during image generation begins testing in the US in October 2026.
OpenAI is also testing Sponsored Agents (ads that open a conversation with a business’s own agent) and an Ads Manager plugin that lets you create and analyze campaigns from inside ChatGPT with AI-suggested copy and imagery, announced in September 2026. We broke that format down in our ChatGPT Sponsored Agents guide.
The partner ecosystem has filled in quickly. HubSpot is the first CRM partner, and Shopify is the first ecommerce partner, with a ChatGPT Ads app. Agency partners include Dentsu, Omnicom, Publicis and WPP, with technology partners such as Adobe, Criteo, Kargo, Pacvue and StackAdapt, per the May 2026 launch.
The build is familiar if you’ve run search before. The parts that trip teams up are technical, not strategic.
If you’d rather run a structured first test than learn the platform slowly in-house, this is exactly the kind of build our team handles as part of PPC management, including ChatGPT Ads tests run alongside your existing Google spend.
We’ll set up and run a ChatGPT Ads test alongside your Google spend, with conversion tracking in place from day one.
The published results are early but specific, and they point in the same direction: new customer acquisition.
Both figures come from OpenAI’s October 2026 update. Earlier, one ecommerce advertiser reported 3x ROAS over 28 days, and a technology partner saw more than 80% of its ad-driven traffic come from new customers.
Read these as prospecting numbers, not retargeting numbers. The pattern across all four is incremental reach, which is the strongest argument for testing if your Google account has plateaued and your CAC keeps climbing on the same pool of queries.
Because incrementality matters more than platform-reported conversions here, OpenAI is running early geo-based experiments with Haus, Measured and WorkMagic, and brand suitability pilots with DoubleVerify and IAS. One expectation to set internally: advertisers get aggregated reporting, not conversation-level data or personal information, as reported at launch.
Since self-serve launched, a theory has spread among advertisers: run ChatGPT Ads and your “organic” ChatGPT traffic goes up, then drops the day you pause the campaign. Some take it further and claim that sponsors get mentioned more often in ChatGPT’s answers, even when users ignore the ad.
The short answer: there’s no evidence that ads change what ChatGPT recommends, and most of the reported lift is an attribution problem. There are, however, a few real indirect effects worth knowing about.
What OpenAI says. OpenAI’s help center is unambiguous: “Ads do not influence ChatGPT’s answers.” Ads “run on separate systems from our chat model,” and “advertisers have no ability to shape, rank, or alter ChatGPT’s responses.” By default, ChatGPT can’t even see the ads shown to a user. Nobody outside OpenAI can audit that, but no public controlled study has shown otherwise.
Why “organic” ChatGPT traffic jumps when ads run. ChatGPT adds utm_source=chatgpt.com to the links in its organic answers, per OpenAI’s Publishers and Developers FAQ. Ad clicks carry a different, OpenAI-specific click identifier called oppref, which is used to match conversions to ad clicks. Standard analytics setups don’t read oppref as a traffic source, so an ad click without your own UTM tags can land in the same chatgpt.com referral bucket as a free citation. Turn the ads off and that “organic” traffic vanishes the same day, which is exactly the pattern people are posting about.
The fix is simple: add your own UTM parameters to every ad URL (OpenAI supports static tracking parameters in Ads Manager) and give paid ChatGPT traffic its own channel in your analytics, ranked above the chatgpt.com referral rule.
Where there’s some truth to it. Ads can still lift your organic results indirectly:
How to test it on your own account. Tag every ad URL, track a fixed set of buyer prompts each month to see whether ChatGPT mentions you, and use geographic targeting to pause ads in one comparable market for three to four weeks while they keep running elsewhere. Then compare tagged sessions, branded search and pipeline between the two. Experiments matter here: when eBay tested paid search with controlled experiments, its measured returns were a fraction of conventional estimates, and brand-keyword ads showed no measurable short-term benefit, because ad clicks and sales were correlated without one causing the other.
If what you really want is more organic ChatGPT mentions, invest in the thing that drives them. ChatGPT recommends the brands it can retrieve and trust: crawlable pages that answer buyers’ questions directly, specific facts and pricing, and credible coverage on third-party sites. That’s what our GEO service is built to do. For Constant Hire, a recruitment agency for ecommerce brands, that work led to more than $70,000 in new business from ChatGPT referrals in four months, without paying for a single ad click. For Buoy, it meant going from absent to ChatGPT’s top recommendation for its main buying prompt in three weeks.
For most B2B and SaaS companies, this is the section that matters.
If your buyers use ChatGPT through Plus, Business or Enterprise accounts, no bid will reach them, because those plans are ad-free. The only way to show up for them is to be cited in the answer itself, which sits above the ad slot and reads as a recommendation rather than a sponsored placement.
That work looks different from classic SEO. ChatGPT builds answers from sources it can retrieve and trust, so the job is to be the clearest, most citable source on the specific questions your buyers ask. That means direct answers near the top of a page, specific numbers with attribution, comparison content that names alternatives honestly, and pages that OAI-SearchBot can crawl. It also means accepting that more of these interactions never produce a click, which we unpacked in our post on zero-click search. For a B2B-specific playbook, see our B2B SaaS ChatGPT guide, and for an example of the results, the Buoy ChatGPT case study.
If you sell to companies on paid ChatGPT plans, organic citations are the channel, not a consolation prize. Our GEO service is built around earning those citations and tracking how often you show up in AI answers for the questions that come before a purchase.
We’ll check which ChatGPT, Perplexity and Google AI answers mention you today, and where competitors show up instead.
Austin Coker is the CEO and founder of 95 Projects. He has worked in SEO since 2020, when he built and sold an ecommerce brand that grew mainly through search. He founded 95 Projects in 2022 to help SaaS, ecommerce, and B2B service companies show up wherever their buyers search, from Google to ChatGPT and AI Overviews.
Have questions about working with us? Book a 30-minute strategy call to discuss your goals and see if we’re a good fit.
No-pressure conversation
Clear next steps (if we’re a fit)
Talk directly to an expert
No. OpenAI limits ads primarily to consumer categories such as household goods, local services, travel, and digital products or education. Financial, health and legal services are reviewed case by case in the US and may need licensing proof, and categories including gambling, political advertising, cannabis and dating are prohibited.
Not on self-serve. The early managed pilot reportedly required about $200,000 upfront, later reduced to $50,000, but the Ads Manager lists no minimum and gives new advertisers $500 in credit after they spend $500.
OpenAI recommends starting with a max CPC bid of $3 to $5. The final price is set in a relevance-weighted second-price auction, so your actual cost depends on how well your ad and landing page match the conversation.
No. Ads appear only for logged-in adults on the Free and Go plans. Plus, Pro, Business and Enterprise accounts don’t see ads, which limits reach for B2B advertisers whose buyers use workplace plans.
Some can, but the policy is written around consumer categories, and buyers on Business or Enterprise plans never see ads. For most B2B and SaaS teams, the better investment right now is earning organic citations in ChatGPT answers.
ChatGPT uses the conversation’s context, your landing page, your ad text and topic-level context hints instead of exact-match keywords, and hints don’t guarantee placement. Product feeds, geographic and platform targeting, custom audiences and negative phrases (for qualifying advertisers) add more control.
Yes. There’s a pixel and a Conversions API, UTM parameters are supported, and measurement and attribution partners include LiveRamp, Tealium, AppsFlyer, Triple Whale and Northbeam. Reporting is aggregated, and spend data can lag seven to eight hours behind other metrics.
Check first that your landing page doesn’t block OAI-AdsBot, which OpenAI requires for landing page review, and allow OAI-SearchBot too. Ads can also be rejected if the creative is misleading or doesn’t match the landing page, and OpenAI may decline advertisers for business or competitive reasons.
No, according to OpenAI, which says ads run on separate systems and advertisers can’t shape, rank or alter ChatGPT’s responses. Most of the reported organic lift comes from untagged ad clicks being counted as chatgpt.com referral traffic. Ads can still help indirectly through brand awareness, and through unblocking OAI-SearchBot, which makes your pages eligible for organic citations.
