What ChatGPT's Ad CTR Data Means for AEO Teams

29 September 2026 · The Cresia team

Search Engine Journal reports that ChatGPT's in-chat ads are clearing a 1% click-through rate in four markets — and the United States isn't one of them. The same report, built on Similarweb panel data, tracks how often the roster of top ChatGPT advertisers turns over from one stretch to the next. Neither number is really about ads. Both are evidence that the answer engine deciding what a user sees is still being tuned in public, market by market, and that should change how AEO and GEO teams plan their next quarter.

Key takeaways

Search Engine Journal's Similarweb-based analysis found ChatGPT ad CTR above 1% in four markets, with the U.S. notably absent — a sign that AI surfaces don't behave uniformly across geographies.
Frequent turnover among top ChatGPT advertisers points to a ranking and auction system that's still shifting, not a stable, learnable playbook.
Paid ad CTR and organic citation rate are different metrics on different mechanics — high performance in one tells you very little about the other.
Answer engines select and cite sources through retrieval and ranking steps that reward clear structure and unambiguous claims over raw domain authority.
The practical fix isn't a new tactic, it's treating AI visibility as market-specific and channel-specific instead of a single global score.
Start measuring now, even roughly, because none of the major AI platforms publish stable historical benchmarks to catch up on later.

What Search Engine Journal's ChatGPT ad CTR report actually found

The report Search Engine Journal published pulls from Similarweb panel data to break ChatGPT ad click-through rates down by market rather than reporting one global figure. Four markets crossed the 1% CTR mark. The U.S. — by far the platform's largest user base — did not. The report also tracked churn in the list of top advertisers, meaning the brands showing up most often in ChatGPT's ad inventory changed noticeably between measurement periods.

Two things are worth sitting with here. First, CTR varying this much by market isn't unusual for a young ad product — Google Ads, Meta and every other platform went through the same uneven early rollout, testing placements and formats against different audiences before anything stabilized. Second, and more relevant to anyone not running ChatGPT ads at all: if the paid side of ChatGPT is this volatile, the organic and cited side is almost certainly moving too. OpenAI doesn't publish a citation-rate benchmark the way Similarweb can approximate an ad CTR, so the ad data is one of the only public proxies we get for "this system is still being calibrated."

Why an ad metric matters if your job is organic AI visibility

It's tempting to read a CTR report and decide it's a media-buying story, not a content or SEO one. That's the wrong read for two reasons.

Advertiser turnover implies signal instability. If the same brands aren't consistently winning the top ad slots, the underlying targeting and relevance model is still being adjusted. Answer engines that surface ads and answer engines that surface cited sources aren't the same system, but they're built by the same company on the same user base, and both are subject to the same pace of change.
Market-level variance is the headline, not the footnote. A U.S.-based team reading "CTR topped 1% in four markets, not the U.S." might shrug. A team running content or campaigns in any of those other markets shouldn't. If ChatGPT's ad engagement differs this much by geography, assume citation and answer-selection behavior differs by geography too — which most GEO measurement setups, built around a single home-market prompt list, currently ignore.

The practical implication: a single "are we visible in ChatGPT" score, checked from one location with one set of prompts, is measuring a system that behaves differently ten miles or ten time zones away.

How AI answer engines actually select and cite sources

Search engines and answer engines share a retrieval step but diverge sharply after that. A classic search ranking pulls from a crawled, indexed corpus and orders results largely on authority, relevance and freshness signals accumulated over time. An answer engine — ChatGPT with browsing, Perplexity, Google's AI Overviews — does something closer to retrieval-augmented generation: it pulls a small set of candidate passages, then a language model decides which ones actually answer the question and how to phrase the synthesis.

That second step is where AEO and GEO diverge from classic SEO in practice:

Passage-level clarity beats page-level authority. A model deciding what to cite is scoring a paragraph, not a domain. A well-known publisher with a vague, hedge-everything paragraph can lose a citation to a smaller site that states the fact plainly.
Structure is a retrieval aid, not just a UX nicety. Headings, defined terms, and direct question-and-answer framing make a passage easier for a retrieval system to match against a query — this is functionally similar to how featured snippets got selected, just with an LLM doing the final pick instead of a snippet algorithm.
Recency and citation format matter differently per engine. Google's own AI Overviews team has been visibly iterating on how citations are displayed — Search Engine Roundtable has tracked cards moving position and loading-state changes within Overviews, which is the same "still being tuned" pattern the ChatGPT ad data shows on the monetization side.
Being a candidate at all is the first gate. None of this matters if a page isn't crawlable, isn't indexed somewhere the model's retrieval layer draws from, or is blocked by the wrong robots directive. That's a technical floor, not an optimization ceiling.

What to change in your AEO/GEO playbook this quarter

None of this calls for a new channel strategy. It calls for treating the current moment as unsettled rather than solved, and building process around that.

Change	Why it follows from the CTR report	Who owns it
Segment AI-visibility tracking by market	Ad CTR variance by geography suggests citation behavior likely varies too	Analytics / SEO lead
Rewrite top pages for passage-level clarity	Retrieval favors clear, quotable paragraphs over hedged authority	Content lead
Re-audit crawlability and indexation quarterly	Being retrievable at all is the gate before ranking or citation matters	Technical SEO
Separate paid AI-surface tests from organic GEO scorecards	CTR and citation rate measure different mechanics — don't conflate them	Media / growth lead
Re-check top queries monthly, not quarterly	Roundtable's tracking of AI Overview citation-card changes shows the surface itself shifts fast	SEO / content ops

A few of these are worth expanding:

Don't wait for a platform to publish a stable API before you start segmenting. Even a manual quarterly spot-check — same prompts, run from a VPN in each priority market — beats a single aggregated number.
Rewriting for passage clarity doesn't mean stripping nuance. It means putting the direct answer in the first sentence of a section and letting caveats follow, so a retrieval system has a clean claim to pull even if the fuller context sits right below it.
Treat AI-surface changes the way you'd treat a search algorithm update: assume your last measurement is already partly stale.

Teams coordinating this across content, media and analytics functions are exactly what a shared operating layer is for — this is the kind of cross-functional tracking work that fits naturally with a platform like Cresia's MediaPilot, which is built around media and AI-search visibility specifically.

A measurement plan for paid and organic AI visibility

The mistake to avoid is building one dashboard that mixes ad performance and citation performance into a single "AI visibility" score. Keep them separate, but run them on the same cadence so you can see when a platform-level shift touches both.

Metric	What it tells you	Where to check it	Cadence
Ad CTR by market (if running ChatGPT ads)	Whether paid placement is landing with the audience you're targeting	Ad platform dashboard	Monthly
Citation frequency for target queries	Whether your content is being pulled into AI-generated answers at all	Manual prompt testing, per market	Monthly
Passage-level match rate	Whether the exact wording cited matches what you intended to be quotable	Manual review against source pages	Quarterly
Crawl/index status for priority URLs	Whether pages are even eligible to be retrieved	Server logs, Search Console-equivalent tooling	Monthly
Referral traffic from AI surfaces	Whether citations are converting into actual visits	Web analytics, segmented by referrer	Monthly

None of these require exotic tooling — a spreadsheet and a recurring calendar hold are enough to start. The point is cadence and separation, not sophistication.

What not to do with this data

A market-level ad CTR report is easy to over-read. A few things to resist:

Don't assume the four markets with strong ChatGPT ad CTR are also the markets with the best organic citation rates — the report doesn't say that, and the mechanics behind the two numbers aren't the same.
Don't treat "not in the U.S." as evidence that ChatGPT ads or ChatGPT search traffic aren't worth U.S. teams' attention. A CTR under 1% in the largest, most competitive market is still a large absolute volume of clicks.
Don't rebuild your whole content calendar around one report. This is one data point in a pattern — Search Engine Roundtable's tracking of AI Overview citation-card placement changes and loading-state tests points the same direction: these are systems still being adjusted, not settled ground truth to reverse-engineer once and leave alone.
Don't let "AI search is unstable" become an excuse to stop measuring. Unstable systems are exactly the ones worth checking on a real cadence, because the cost of missing a shift is higher, not lower.

The bigger pattern behind the ad numbers

Zoom out and the ChatGPT CTR report fits a pattern showing up across the AI search space this year: platforms visibly adjusting core mechanics in production. Search Engine Journal separately covered CallRail's move to make ChatGPT ad performance measurable for smaller advertisers — itself a sign that measurement infrastructure for this channel is still being built, not mature. Google has been testing where AI Overview citation cards sit on the page and how the "loading" state renders, changes Search Engine Roundtable has tracked in near real time. None of this is a finished product being incrementally polished. It's infrastructure being assembled while millions of people already use it daily.

For a marketing team, the operational takeaway is the same one that applies to any channel built on a platform you don't control: build measurement before you build conviction. Teams that already have a habit of tracking search and campaign performance across fragmented surfaces — the kind of work covered under marketing operations — are better positioned to absorb this than teams treating AI search as a one-time audit. And because paid and organic AI visibility now sit on the same platforms but move independently, media and analytics functions need a shared view rather than two separate reports that never get compared, which is the coordination problem solutions built for media teams and analytics teams are meant to close.

The honest summary: nobody, including OpenAI's own advertisers, has this fully figured out yet. That's not a reason to wait. It's the reason to start measuring now, while the baseline is still cheap to establish.

Frequently asked questions

Does a strong ChatGPT ad CTR in a market mean organic citations are also strong there?

Not necessarily. Ad CTR measures whether a paid placement got clicked; citation rate measures whether a model chose to quote or reference a source while generating an answer. They're produced by different parts of the system, so a market performing well on one metric can look completely different on the other. Track them separately.

Which four markets did Search Engine Journal report crossing 1% CTR?

Search Engine Journal's report breaks this down by market using Similarweb panel data; check the source report directly for the specific market list, since that detail is best read from the original breakdown rather than summarized secondhand.

Should a team without a ChatGPT ad budget care about this report?

Yes, for what it implies rather than what it states. Advertiser turnover and market-level CTR variance both suggest OpenAI's targeting and relevance systems are still being tuned, which is relevant to anyone trying to get cited organically, not just anyone buying placements.

How is this different from the AI Overviews citation-card testing Search Engine Roundtable has covered?

Different platform, same underlying pattern: a company actively adjusting how sources are surfaced and displayed in an AI-generated answer. Google's changes are about citation card placement and loading behavior inside Search; OpenAI's ad data is about paid placement performance inside ChatGPT. Neither is finished settling.

What's the fastest way to start measuring AI-surface visibility if we have nothing in place?

Pick 10-15 real customer questions, run them manually against ChatGPT, Perplexity and Google AI Overviews from your top two or three markets, and log whether and how your brand gets cited. Repeat monthly. It's not sophisticated, but it beats having zero baseline when the next platform change lands.

Sources

https://www.searchenginejournal.com/chatgpt-ad-ctr-markets-report/591384/
https://www.searchenginejournal.com/callrail-makes-chatgpt-ads-measurable-for-smbs-and-marketing-agencies-spn/589843/
https://www.seroundtable.com/google-ai-overviews-citations-cards-bottom-42177.html
https://www.conductor.com/academy/aeo-geo-benchmarks-report/