ChatGPT Ads CTR and Benchmarks: What to Expect

ChatGPT Ads CTR runs well below Google Search in early advertiser reports, with one cited Adthena figure near 0.91 percent versus a multiple-times-higher Google Search benchmark. That gap is structural, not a failure. People often continue the conversation instead of clicking out, so a click is a higher-intent and rarer action here. CTR is the wrong headline metric for this channel. Measure cost per qualified outcome through the OpenAI pixel, Conversions API, and UTMs instead.

What Is a Normal CTR on ChatGPT Ads?

There is no official benchmark yet, because OpenAI has not published cross-advertiser CTR data. What exists is early and advertiser-reported.

The most-cited early figure puts ChatGPT Ads CTR near 0.91 percent, reported via Adthena. That sits well below typical Google Search click-through rates, which run several times higher depending on the category and ad position.

Treat that 0.91 percent as directional, not as a target you should hold your account to. It comes from a thin slice of early advertisers in the first months of a brand-new format. Your own CTR will depend on your offer, your ad copy, the context where your ad appears, and how relevant the placement is to the live conversation. A figure pulled from someone else's account in a different category tells you very little about what yours will do.

The honest answer to "what is normal" right now is that nobody has enough data to say. Anyone quoting a precise ChatGPT Ads CTR benchmark as established fact is overstating what the channel has produced so far.

Why Is ChatGPT Ads CTR Lower Than Google Search?

The gap comes from where the ad sits and what the user is doing when they see it.

On Google Search, the person has typed a query and is looking at a list of links. Clicking a link is the entire point of the interaction. The user arrived expecting to leave the results page and go somewhere. Ads sit inside that click-to-leave behavior, so click-through rates are naturally higher.

On ChatGPT, the person is in a conversation. The ad appears below a relevant response, clearly labeled and separated from the answer. The default behavior is to keep reading, ask a follow-up, or refine the request, not to leave the conversation. A click out is a break in what the user came to do, which makes it a rarer and higher-intent action.

Targeting is also contextual rather than keyword-exact. The system weighs context hints, the landing page, the ad title, and the ad copy, then selects primarily on relevance to the live conversation. Context hints are not exact-match keywords and do not guarantee delivery, so the audience is matched on conversational relevance rather than a typed commercial query. None of this means the channel underperforms. It means a click here is worth more per unit than a click on search, and the raw CTR number reflects a different user moment.

Why Is CTR the Wrong Headline Metric for This Channel?

CTR measures one thing: how often people click. It says nothing about whether those clicks become customers, which is the only outcome that pays for the spend.

On a channel where clicking out interrupts the user's task, optimizing for raw CTR pushes you toward the wrong behavior. You would write baitier ad copy, chase clicks from people who are not ready to act, and inflate a vanity number while your cost per actual outcome gets worse. A higher CTR that delivers lower-quality clicks is a worse result, not a better one.

CTR also looks artificially bad here precisely because the channel works differently. If you judge ChatGPT Ads by a search-style CTR benchmark, you will conclude the channel is failing when it may be delivering qualified outcomes at an acceptable cost. The number is low by design, not by malfunction.

The metric that matters is what each qualified outcome costs you. A placement with a 0.7 percent CTR that produces leads at a price you can afford beats a 2 percent CTR that produces clicks who never convert. Headline CTR is a distraction from that math.

What Should You Measure Instead of CTR?

Measure cost per qualified outcome, and build the tracking to see it before you spend.

OpenAI's measurement stack is the pixel, the Conversions API, and UTMs. The native reporting gives you impressions, clicks, spend, CTR, average CPC, and average CPM. Those platform numbers tell you what the channel charged you. They do not tell you what you got for it. For that, you connect the click to what happens on your site: a form fill, a booking, a qualified lead, or a sale.

The sequence that matters: define the outcome that represents real business value, install the pixel and Conversions API so that outcome is tracked, tag your destination URLs with UTMs so ChatGPT-sourced traffic is identifiable in your analytics, then judge the channel on cost per qualified outcome rather than CTR or even raw conversions. A high raw conversion count made of low-quality leads is its own trap.

This is the same discipline that separates a measurable Google Ads account from a blind one. The platform is new, but the principle is not: the number that decides whether a channel earns more budget is what a real outcome costs, not how many people clicked. We cover the full setup in the post on ChatGPT Ads conversion tracking.

What Benchmarks Exist So Far, and How Reliable Are They?

A few directional figures exist. None are official OpenAI benchmarks, and all should be read as early-stage signals.

On cost, CPC bidding was added April 2026, with a recommended starting bid of $3 to $5 per click. Bids under $3 may fail to clear delivery, so treat $3 to $5 as a working floor. CPM launched at a flat $60 and category rates have since settled to roughly $25 to $60. Cost-per-action bidding went live on June 5, 2026. On engagement, the early CTR figure near 0.91 percent comes via Adthena. These are starting points for planning, not guarantees.

The reliability problem is sample size and age. The format launched in February 2026, the open Ads Manager arrived in May, and every number comes from a small population of early advertisers in their first weeks of spend. Reported figures are also self-selected: advertisers who share numbers are not a representative cross-section of all accounts. OpenAI has not released aggregate cross-advertiser CTR, CPC, or conversion benchmarks.

Use these figures to set a planning range and a CPC floor, then replace them with your own account data as fast as you can. Your real benchmark is your last 30 days, not someone else's screenshot. We go deeper on pricing in the post on ChatGPT Ads cost.

How Do You Improve Performance Without Chasing CTR?

You improve performance by improving relevance and the post-click experience, not by gaming the click rate.

The auction is relevance-weighted and, as trade-press reporting describes it, second-price style, which means a sharp, relevant ad can win without the highest bid. The system weighs your context hints, landing page, ad title, and ad copy together and selects on relevance to the live conversation. So the levers are: tighten your context hints to the conversations where your offer genuinely fits, write ad copy that matches the intent of those conversations, and point clicks to a landing page that delivers exactly what the ad promised. Relevance lowers what you pay and raises the quality of who clicks.

After the click, the work is the same discipline any paid channel demands. Make sure the landing page loads fast, matches the ad message, and makes the next step obvious. A relevant click that hits a slow or mismatched page is wasted spend regardless of how good the CTR looked.

Then optimize against cost per qualified outcome over a meaningful window, not day-to-day CTR swings. Give the placement enough time and spend to produce outcome data, cut what does not produce qualified outcomes at an acceptable cost, and put budget behind what does. Chasing CTR optimizes for the wrong thing. Chasing cost per qualified outcome optimizes for the business.

How Do You Set Expectations With Stakeholders?

Set the frame before anyone sees the dashboard, because the CTR column will look alarming next to Google Search if you do not.

Tell stakeholders three things up front. First, CTR on this channel runs well below search by design, because users continue the conversation instead of clicking out, so a low CTR is expected and not a sign of failure. Second, the channel is new, the benchmarks are early and advertiser-reported, and the real measure of success is cost per qualified outcome, which the platform CTR number does not show. Third, results will be judged over a window long enough to gather outcome data, not on the first week of clicks.

Put the right number in front of them. Report cost per qualified outcome, the volume of qualified outcomes, and spend, with CTR and CPC shown as context rather than the headline. If the only number a stakeholder sees is a 0.9 percent CTR, they will draw the wrong conclusion no matter how well the channel is performing on outcomes.

This is also a judgment call about whether the channel fits at all, which we walk through in the post on whether ChatGPT Ads are worth it. Setting honest expectations early is what keeps a working channel from getting killed over a metric that was never the right one to watch.

Sources

• Campaign US: underwhelming early returns on ChatGPT Ads (the 0.91 percent versus 6.4 percent CTR figure against a higher Google Search benchmark)
• OpenAI: our approach to advertising and expanding access (measurement stack, formats, no published cross-advertiser benchmarks)
• OpenAI: new ways to buy ChatGPT Ads (self-serve, CPC, no minimum, and partners)