ChatGPT Ads for B2B SaaS: CTR by Market, Real Costs, and When to Earn Citations Instead

Since OpenAI started testing ads in ChatGPT on 9 February 2026, the question I hear most from B2B founders has changed. It used to be "how do we get into ChatGPT's answers?" Now it is "can we just buy our way in?"

That is a fair question. ChatGPT Ads reached a $1 billion annualized revenue run rate in under 200 days, and B2B SaaS brands were among the first to show up. In June, eight of the top ten advertisers by global impression share were B2B SaaS or productivity brands.

But one line in OpenAI's own help center changes the whole calculation. The ad appears below the end of the answer, and OpenAI says ads run on separate systems from the answer itself. Below are the numbers, market by market. I price a B2B click all the way down to a qualified lead, and show where the budget earns its keep.

1. Are ChatGPT Ads Worth It for B2B SaaS Right Now?

ChatGPT ads are worth a small, measured test for B2B SaaS, not a budget shift. Reported CPC bids run $3 to $5, and US click-through rates sit near 0.73%. One documented B2B test turned 739 clicks into zero qualified leads. The ad also sits below ChatGPT's answer, not inside it. MaximusLabs treats ads as a test on top of earned citations, because the recommendation inside the answer is where B2B buyers form their shortlist.

🧭 How B2B demand used to be bought

For fifteen years, B2B SaaS teams bought demand in two places. You bid on high-intent keywords in Google, and you targeted job titles and company sizes on LinkedIn. Both channels let you pay to stand in front of a buyer at the moment of research.

That model had a cost we all accepted. Clicks on software terms got expensive, and LinkedIn charged a premium for firmographic precision. But the logic held: pay for position, win the visit, convert the visit.

🔄 What changed when buyers moved into the chat

Buyers now start somewhere else. G2 surveyed 1,076 B2B decision-makers in 2026. It found that 51% of software buyers now start their research in AI chatbots, up from 29% in April 2025. The same survey found 69% ended up choosing a different vendor than they first planned.

So the shortlist now forms inside an answer, not on a results page. When a buyer asks ChatGPT for the best tools in your category, ChatGPT names a handful of brands and explains why. The ad, when one appears, arrives after that decision has already been shaped.

🎯 How MaximusLabs approaches the paid versus earned question

At MaximusLabs we split this work into two pillars. Owned AEO makes your own content the answer engines extract. Earned AEO gets you named in the third-party sources ChatGPT trusts, such as review sites, Reddit threads, and comparison articles. ChatGPT Ads sit outside both pillars, so we run them as a bounded experiment with a pipeline target, never as a substitute for being cited.

2. How Do ChatGPT Ads Actually Work in 2026?

ChatGPT ads are sponsored units that appear below the end of a ChatGPT response, clearly labeled and visually separated from the answer. They show only to Free, Go, and logged-out users. Plus, Pro, Business, Enterprise, and Edu accounts see no ads. OpenAI says ads run on separate systems from the chat model, and advertisers cannot shape, rank, or alter answers. Buying is self-serve through Ads Manager, with CPC, CPM, and conversion bidding.

🧱 Where the ad appears and who sees it

OpenAI's help center describes the placement plainly: "Ads can appear below the end of a response. Ads are clearly labeled as sponsored and visually separated from ChatGPT's response." That is the single most important fact for a B2B marketer. The answer comes first, and the ad follows it.

Eligibility is narrow by design. Ads may appear on the Free and Go plans and in logged-out sessions. Accounts identified as belonging to users under 18 never see ads. Ads are also kept away from sensitive topics, including personal health, mental health, and politics.

🧮 How OpenAI picks which ad to show

When several ads qualify, OpenAI weighs "a combination of factors, including relevance and advertiser bids." The relevance signals include the current conversation, the ad's landing page, its title and copy, and "advertiser-provided context hints." Where personalization is on, past chats and memory can also count. Personalized ads are not initially available in the European Economic Area or Switzerland.

There are two formats worth knowing. The original unit is a single sponsored card with a headline, description, image, and link. In August, OpenAI added product carousels that pull from retailer feeds, which matter more to ecommerce than to SaaS.

💳 How buying works today

Buying opened up fast. At the February pilot, PPC Land reported pricing of a $60 CPM with a $200,000 minimum commitment. By 5 May, OpenAI had launched a self-serve Ads Manager in the US with cost-per-click bidding and no minimum spend. By late August, it had added conversion optimization, geo-targeting, custom audiences, the OpenAI Pixel, and a Conversions API.

Access now runs through three doors. There is OpenAI's Ads Solutions team, agency partners such as dentsu, Havas, Omnicom, Publicis, and WPP, and self-serve Ads Manager. If you want the organic side of this platform first, our ChatGPT search optimization guide covers how ChatGPT retrieves and cites sources.

3. Where Are ChatGPT Ads Live, and What CTR Does Each Market Get?

ChatGPT Ads are available in over 60 countries after the 23 September 2026 expansion to Southeast Asia and Taiwan. Click-through rates vary sharply by market. Graphite and Similarweb panel data show Brazil at 1.91%, Mexico at 1.55%, Japan at 1.53%, and South Korea at 1.21%. The US sits at 0.73%, the UK at 0.74%, and Canada at 0.75%. Only those four newer markets clear 1%.

🗓️ The rollout, date by date

The expansion moved in waves. Most dates below come from OpenAI's announcements; the UK go-live is from Similarweb and the India launch from TechCrunch.

📊 CTR by market, and how to read it

Graphite's "State of ChatGPT Ads" report splits CTR by market. It is built on Similarweb panel data and was covered by Search Engine Journal on 29 September.

Read these numbers with care. The CTR counts clicks per ad-bearing exchange, it is pooled across all advertisers, and it is an all-time rate through 16 September. Similarweb shares percentages but not panel volumes. Markets that launched later have shorter windows and newer auctions, so part of the gap may simply be novelty.

🌍 What the spread means for a B2B team

For a US-first SaaS company, the honest benchmark is about 0.7%. For context, Similarweb cited a 0.35% display benchmark in May, from a different dataset and period. So the unit is not weak. It is simply not a high-intent search click.

The Brazil and Japan numbers are interesting if you sell there, but I would not build a plan on them yet. If your growth depends on the US, UK, or Canada, plan around the 0.73% to 0.75% band. Our guide to AI search click-through rates explains why clicks from AI surfaces behave differently from classic search.

4. What Does a ChatGPT Ad Really Cost a B2B SaaS Team per Qualified Lead?

Reported ChatGPT CPC bids run $3 to $5 per click (The Next Web, April 2026). B2B SaaS should judge the channel on cost per qualified lead, which can run into the thousands. Workshop Digital's documented B2B test spent $2,319 for 739 clicks at a $3.14 CPC and produced zero qualified leads. MaximusLabs prices every channel by pipeline, because a cheap click that never becomes an opportunity is still wasted spend.

💸 How B2B teams used to judge a new channel

The traditional way to judge a new paid channel was CPC and CTR. If clicks were cheaper than Google and the click rate looked healthy, the channel passed. Leadership saw a dashboard of efficient clicks and approved more budget.

That habit hid a real cost. B2B SaaS has long sales cycles, several decision-makers, and contract values that justify a careful purchase. A click that arrives early in research rarely converts on the first visit, so CPC alone overstates how well a channel works.

🔬 What the first real B2B test showed

Workshop Digital published the clearest B2B test so far in August. It spent $2,319, earned 69,630 impressions and 739 clicks, and averaged a 1.1% CTR at a $3.14 CPC. Keyword-style context hints did best, at a 1.2% CTR and a $3.13 CPC. Financial services ran slightly worse, at 0.9% and $3.44.

The result that matters came last: no qualified lead. Workshop Digital was careful to say this is one advertiser at one moment in a young product. I agree. But it matches the pattern Graphite flagged in its own guidance. The top ten advertisers "all sell something buyable in one sitting," so "test long sales cycles with caution."

🧮 The cost per qualified lead math

Here is the math I run with founders. The CPC comes from the Workshop Digital test. The two conversion rates are illustrative assumptions, so swap in your own.

The spread is the point. The same $3.14 click can cost $349 or $3,140 per sales-qualified lead, depending on what happens after the click. Most of that range is decided by your landing page and your offer, not by ChatGPT.

📈 How MaximusLabs prices a channel

MaximusLabs measures every channel against pipeline, not clicks. For AI search that means tracking which prompts name you and which of them carry buying intent. It also means counting how many opportunities cite ChatGPT as a source. If you run a ChatGPT Ads test, hold it to the same bar. Set the cost per SQL your Google and LinkedIn programs already hit, and make ChatGPT beat it before you scale. Our GEO budget benchmark shows how B2B teams are splitting AI search budgets in 2026.

5. Can ChatGPT Ads Target Your ICP the Way LinkedIn or Google Can?

No. ChatGPT Ads target the context of a conversation, not a job title or company. OpenAI matches ads to what the user is discussing, plus context hints, geography, and custom audiences. There is no firmographic layer like LinkedIn's title and company-size filters. MaximusLabs works the same gap from the organic side, mapping the prompts each client's buyers ask so the brand is named in those answers.

🏢 How ICP targeting used to work

LinkedIn built a premium business on one promise. Show this ad to VPs of Marketing at SaaS companies with 200 to 1,000 employees. Google Ads added intent through keywords. Together they let a B2B team describe its ideal customer profile (ICP) and pay to reach only those people.

That precision was expensive, but it was predictable. You knew who saw the ad, and you could defend the spend in a board meeting.

🔄 What changed with conversational targeting

ChatGPT Ads work differently. OpenAI says its system "uses the context of the current conversation to show an ad relevant to what someone is exploring." Depending on country and settings, it may also use the user's broader ChatGPT history. Advertisers add context hints, geo-targeting, platform targeting, and custom audiences.

What you cannot do is say "show this only to CFOs." The conversation is the targeting. A question about "SOC 2 automation for a 50-person startup" is a strong signal. But you are inferring the buyer from the question, not selecting the buyer. Workshop Digital's test suggests keyword-style hints outperform conversational phrasing, which fits that logic.

🗺️ How MaximusLabs maps the prompts your buyers ask

The question is the new targeting unit, so MaximusLabs models the questions each client's buyers ask. We build a per-client intelligence graph of the product, ICP, competitors, and pains. Then we turn it into a prompt set that reflects real buying situations. We track those prompts daily across ChatGPT, Perplexity, Gemini, Claude, and Google AI Overviews.

That same prompt map makes a ChatGPT Ads test sharper. The prompts where you are already cited become your control group. The high-intent prompts where you are missing become your first context hints. If you want the method behind this, our AEO strategy planning guide walks through it.

6. Do ChatGPT Ads Help Your Brand Get Cited Inside ChatGPT's Answers?

No. Buying ChatGPT ads does not get your brand cited or recommended inside ChatGPT's answers. OpenAI states that ads "run on separate systems from our chat model" and that advertisers cannot "shape, rank, or alter" responses. The answer and the ad are two different surfaces. MaximusLabs earns the in-answer mention through owned content and third-party citations, because that is the lever that changes what ChatGPT says.

📰 How visibility used to be bought

In classic search, paid and organic sat side by side on the same page. You could buy the top ad slot and rank in the organic results below it. Many B2B teams treated the two as one budget, because the searcher saw both at once and chose.

That habit carries a hidden assumption into ChatGPT: that paying for the ad also buys some presence in the answer. It does not.

🧱 What OpenAI's separation rule means

OpenAI's help center is explicit: "Ads do not influence ChatGPT's answers." It adds that seeing an ad "doesn't mean OpenAI endorses or recommends the advertiser." The persuasive work happens in the answer, where ChatGPT names brands and gives reasons. The ad arrives after the reasoning is done.

Now add what the answer does to clicks. In the NYT v. OpenAI case, a publishers' brief filed on 17 September quotes a 2023 note from an OpenAI engineer. It reads: "no matter how prominently we show the links, users won't click." That is the plaintiffs' framing of discovery material, and OpenAI disputes the case. It also fits the zero-click pattern behind the search referral traffic decline: the answer satisfies the buyer, and the click is optional.

🏗️ How MaximusLabs earns the in-answer mention

Getting named inside the answer is the work of the two pillars of AEO. Owned AEO structures your pages so ChatGPT can lift a clean, self-contained answer, with the brand, the claim, and the reason in one passage. Earned AEO places your brand in the sources ChatGPT trusts for your category: review platforms, community threads, and comparison articles.

This is why I say AEO is brand building evolved. An ad rents a slot below the answer for one session. A citation puts ChatGPT's own credibility behind your name, each time ChatGPT names you for that question. Our citation acquisition tactics guide lists the sources ChatGPT cites most for B2B categories.

✅ The proof that earned visibility compounds

7. Do Your B2B Buyers Even See ChatGPT Ads?

Many B2B buyers never see ChatGPT ads, because ads appear only on Free, Go, and logged-out sessions. Anyone researching on a Plus, Pro, Business, Enterprise, or Edu account sees an ad-free answer. Graphite and Similarweb report that 43.4% of ad-eligible US users had seen at least one ad by September 2026. For a buyer on a company ChatGPT seat, the organic answer is the only thing on the screen.

👥 The eligibility gap in numbers

ChatGPT now has 1.2 billion weekly users, according to OpenAI's DevDay recap on 29 September. That is a huge audience, but the ad-eligible slice is smaller than it looks. Paid consumer plans and all workplace plans are excluded by design. A Free user can even opt out of ads in exchange for lower usage limits.

Reach inside the eligible group is still building. Graphite's data shows how many eligible users had seen at least one ad. It was 43.4% in the US, 45.7% in the UK, 42.9% in Canada, and 35.9% in Japan.

🧑‍💼 Why this matters more for B2B than B2C

Think about how a Head of RevOps actually researches a tool. Many do it at work. If their company provides a ChatGPT Business or Enterprise seat, none of those sessions carry ads. The same buyer on a personal Free account at night would see them.

There is no public figure yet for how much B2B software research happens on paid or workplace seats. So I would not guess. I would ask. Add "Which AI tool did you use, and on what plan?" to your win-loss interviews. Within a quarter you will know how much of your market the ads can reach.

🧾 The practical takeaway

Treat ChatGPT Ads as a channel that reaches part of your buyers, some of the time. Treat the organic answer as the channel that reaches all of them, every time they ask. MaximusLabs covers the second through ChatGPT optimization, so the buyer on a Business seat still hears your name.

8. Should You Buy ChatGPT Ads or Earn Citations First?

For most B2B SaaS companies, earn citations first and test ads second. Earned citations reach every ChatGPT user, including ad-free Business and Enterprise seats, and they keep working without a per-click cost. Ads make sense as a bounded test once your brand is already named in the answers for your core prompts. MaximusLabs sequences the two by stage, so ad spend amplifies visibility you already own instead of renting visibility you lack.

⚖️ How teams usually decide

When a new ad platform appears, B2B teams tend to jump in early. Being first into a new auction often means cheap clicks, and nobody wants to explain to the board why a competitor showed up first. Similarweb's data shows Monday.com alone held 5.05% of global impressions in June. Early movers arrived fast.

The cost of that instinct is sequencing. If ChatGPT's answer names three competitors and not you, an ad below it asks the buyer to overrule the recommendation they just read. That is a hard sell at any CPC.

🔄 What changed: the answer sets the shortlist

In the old search results page, the ad and the organic listing competed on equal footing. In ChatGPT, the answer comes first and carries the platform's credibility. This is Ries and Trout's Positioning playing out in a chat window: the answer takes the position in the buyer's mind first, and the ad fights for what is left. G2 found 69% of B2B buyers chose a different vendor than planned after AI research. And 33% bought from a vendor they had not known before. The answer is now doing the persuading.

🧭 How MaximusLabs sequences ads and citations by stage

Here is the decision rule I recommend. It depends on whether ChatGPT already names you for the prompts that matter.

MaximusLabs publishes its pricing: credit-based plans start at $2,999 a month. The first article can ship as early as day 4, and initial citation movement typically appears in 4 to 8 weeks.

9. How Do You Run a ChatGPT Ads Test Without Wasting Budget?

Run a ChatGPT Ads test for four to six weeks with a fixed budget and a pipeline target set before launch. Use keyword-style context hints and a dedicated landing page. Track conversions with the OpenAI Pixel or Conversions API, tag every click with UTMs, and judge the test on cost per sales-qualified lead, not CTR. Graphite's own guidance is to "keep spend experimental" while the advertiser base settles.

🧪 A seven-step test plan

1. Set the pipeline target first. Decide the cost per SQL that would make the channel worth scaling. Use your Google and LinkedIn numbers as the bar.
2. Cap the budget. Workshop Digital learned a lot from $2,319. A similar fixed amount is enough for a first read.
3. Start with keyword-style context hints. In Workshop Digital's test they beat conversational phrasing, at a 1.2% CTR and a $3.13 CPC.
4. Pick prompts where you are already cited. Your ad then reinforces the answer instead of contradicting it.
5. Build a dedicated landing page. Match the question the buyer just asked. Offer a next step that fits early research, such as a benchmark or a calculator.
6. Install measurement before launch. Use the OpenAI Pixel or Conversions API, UTM parameters, and a CRM field for "AI assistant" as a source.
7. Review weekly, decide at week six. Graphite found the first few turns of a conversation get the highest CTR and impression share. Watch where in the conversation your ad lands.

🧯 Common mistakes to avoid

• Judging the test on CTR alone. A 1% CTR with no pipeline is a failed test.
• Sending clicks to a generic homepage. The buyer arrives mid-question and needs the answer continued.
• Reusing search ad copy. Graphite found top advertisers write 98% new copy for ChatGPT.
• Running financial-services ads outside the US without checking policy. OpenAI's ad policy says these are "generally prohibited" there.

MaximusLabs can run the organic prompt map alongside your test. You then see whether ads lift branded search and citations, or simply buy clicks you would have earned.

10. How Do You Measure Pipeline From ChatGPT Ads and AI Citations Together?

Measure ChatGPT Ads with the OpenAI Pixel, Conversions API, and UTMs. Measure AI citations with daily prompt tracking plus self-reported attribution in your CRM. Ads give you aggregated clicks and conversions, while citations often produce no click at all. MaximusLabs pairs daily AI visibility tracking with self-reported attribution, so AI-sourced pipeline shows up in the CRM.

📉 How attribution used to work

B2B attribution was built on the click. A buyer clicked an ad or a result, a cookie followed them, and the CRM credited the channel when they converted. Every tool in the stack assumed the visit was the unit of measurement.

That model breaks in AI search. If the answer persuades the buyer and no click happens, last-click attribution records nothing. The channel that did the work looks like it did nothing.

🔄 What changed: two surfaces, two kinds of evidence

ChatGPT Ads can be measured the old way. OpenAI offers the Pixel, a Conversions API, and third-party measurement integrations, and advertisers see aggregates such as total views and clicks. CallRail added ChatGPT ad attribution for calls, texts, and forms in June.

The citation inside the answer needs different evidence. You measure whether ChatGPT names you, for which prompts, and against which competitors. Then you connect it to pipeline through what buyers tell you.

📡 How MaximusLabs connects visibility to revenue

MaximusLabs tracks AI visibility daily across ChatGPT, Perplexity, Gemini, Claude, and Google AI Overviews. For each client we report share of voice on the prompts that map to buying stages. Share of voice means how often the brand is named compared with competitors. Then tie it to revenue with a "how did you hear about us" field that lists AI assistants as options. I recommend making it required.

Our GEO metrics and KPIs guide lists the full measurement set.

What I'm Thinking About Next

My current thinking is that ChatGPT Ads will become a real B2B channel. But not in the shape most teams expect. The format that interests me most is one OpenAI has only described so far. It is an ad you can message, where the buyer asks the advertiser questions inside the chat. That turns the ad into a sales conversation, and B2B is built on sales conversations.

I might be wrong about the timing. But I am fairly confident about the order. The brands that win the ad slot will be the ones ChatGPT already names in the answer above it. Earn the answer first, then rent the space below it. That is the sequence I am recommending to every founder who asks me this month.

References

1. OpenAI, "Our approach to advertising and expanding access to ChatGPT," 16 January 2026. https://openai.com/index/our-approach-to-advertising-and-expanding-access/
2. OpenAI, "Testing ads in ChatGPT," 9 February 2026, with updates on 26 March, 7 May and 11 August 2026. https://openai.com/index/testing-ads-in-chatgpt/
3. OpenAI, "New ways to buy ChatGPT ads," 5 May 2026. https://openai.com/index/new-ways-to-buy-chatgpt-ads/
4. OpenAI, "ChatGPT Ads expands across Europe," 18 August 2026. https://openai.com/index/chatgpt-ads-expands-across-europe/
5. OpenAI, "A milestone in expanding access to AI," 31 August 2026. https://openai.com/index/expanding-access-to-ai-with-chatgpt-ads/
6. OpenAI, "ChatGPT Ads expands to Southeast Asia and Taiwan," 23 September 2026. https://openai.com/index/chatgpt-ads-expands-southeast-asia-taiwan/
7. OpenAI Help Center, "Ads in ChatGPT," accessed 30 September 2026. https://help.openai.com/en/articles/20001047-ads-in-chatgpt
8. OpenAI, "Ad policies," accessed 30 September 2026. https://openai.com/policies/ad-policies/
9. OpenAI, "DevDay 2026 Recap," 29 September 2026. https://openai.com/index/devday-2026-recap/
10. Graphite, "The State of ChatGPT Ads" (Similarweb panel data), 28 September 2026. https://graphite.io/research/state-of-chatgpt-ads
11. Search Engine Journal, "ChatGPT Ad CTR Topped 1% In 4 Markets, Not In The U.S.: Report," 29 September 2026. https://www.searchenginejournal.com/chatgpt-ad-ctr-markets-report/591384/
12. Similarweb, "ChatGPT ad stats," 15 June 2026. https://www.similarweb.com/blog/marketing/advertising/chatgpt-ad-stats/
13. Search Engine Roundtable, "ChatGPT ads click-through rates" (Similarweb data), 5 May 2026. https://www.seroundtable.com/chatgpt-ads-click-through-rates-41267.html
14. Workshop Digital, "ChatGPT Ads: What We Learned From Our First B2B Test," August 2026. https://www.workshopdigital.com/blog/chatgpt-ads-for-b2b/
15. G2, "The Answer Economy: How AI Search Is Rewiring B2B Software Buying," 15 April 2026. https://www.prnewswire.com/news-releases/new-g2-research-half-of-b2b-software-buyers-now-start-their-research-with-ai-chatbots-302742807.html
16. Search Engine Journal, "Court Filing Quotes OpenAI Engineer: Users 'Won't Click' Links," 28 September 2026. https://www.searchenginejournal.com/openai-engineer-users-wont-click-links-court-filing/591303/
17. PPC Land, "ChatGPT ads finally leave the US," 7 May 2026. https://ppc.land/chatgpt-ads-finally-leave-the-us-uk-japan-korea-brazil-and-mexico-next/
18. The Next Web, "OpenAI launches CPC ads in ChatGPT," April 2026. https://thenextweb.com/news/openai-chatgpt-cpc-ads-launch
19. Digiday, "OpenAI brings product carousels to ChatGPT ads," 6 August 2026. https://digiday.com/marketing/openai-brings-product-carousels-to-chatgpt-ads/
20. CallRail, "CallRail makes ChatGPT ads measurable for SMBs and marketing agencies," 10 June 2026. https://www.prnewswire.com/news-releases/callrail-makes-chatgpt-ads-measurable-for-smbs-and-marketing-agencies-302796092.html