Channel sheet · CH-20 · gain 4 min · logged November 16, 2026

AI MarketingDirect input

ChatGPT Ads Hit $1 Billion Run Rate, But Early Data Shows Weaker Returns

OpenAI's ads product reached a $1 billion annualized run rate, but a study of 102 websites shows ChatGPT ad visits returning about one-third of paid search revenue per session.

By Nathan Brooks4 min read815 words

Signal notes

  1. OpenAI Ads hit a $1 billion annualized run rate by Aug. 31, with 100,000+ advertisers since February.
  2. U.S. paid search sessions generated $2.45 revenue versus $0.89 for ChatGPT ad sessions.
  3. 84 of 87 advertisers using both channels earned less revenue per session from ChatGPT ads.
  4. Frankfurt School figures of $0.52 vs $3.40 per ad dollar rest on one agency's $1.72 ChatGPT CPC and a $0.72 median search CPC.
  5. StubGroup case: ChatGPT cost per lead $92.21 vs Google's $157.57, but Google leads converted to customers at a higher rate.
ChatGPT Ads Vs. Google Search Ads: What Early Data Shows
Input monitorChatGPT Ads Vs. Google Search Ads: What Early Data Shows — Pexels / pixabay

OpenAI's ads product passed a $1 billion annualized run rate by Aug. 31, with more than 100,000 advertisers running campaigns since the February launch, according to an Oct. 21 OpenAI engineering post. The number sounds large. Against Alphabet's $63.35 billion in Search & other revenue for the third quarter, up 12%, it is roughly one-sixtieth — and no data exists showing how many of those dollars came out of Google Search budgets.

What does the early evidence actually show?

A working paper by researchers Kaiser and Schulze, based on 102 websites with transaction and revenue tracking donated by Grips Intelligence, found that a ChatGPT ad visit generated about one-third the revenue of a paid search visit in U.S. traffic. Paid search produced $2.45 in revenue per session versus $0.89 for ChatGPT ad sessions. Transactions per session ran 1.17% against 0.37%. The pooled data covers 60 weeks through Aug. 22, with up to 28 weeks of ChatGPT ad sessions for U.S. traffic.

The model controlled for website, device type, and week, and results held after removing individual advertisers. Of 87 advertisers using both channels in U.S. traffic, 84 produced less revenue per session from ChatGPT ads. The authors call the work "descriptive rather than causal" — they do not claim switching matched budgets between platforms would help or hurt any specific advertiser.

Which figures were measured, and which were assembled?

Here the story gets murkier. A press release from one author's institution, the Frankfurt School, stated that advertisers receive $0.52 back per dollar spent on ChatGPT Ads versus $3.40 for Google Search. Only the per-session revenue figures were measured. The per-dollar figures combine them with two external prices: a $1.72 ChatGPT cost per click that agency Opascope published in June, and a $0.72 median search CPC from 16 mostly U.S. accounts. Those prices were never matched to the revenue each site generated.

The $1.72 figure covers one account, about $60,000 in spend over 15 days, on campaigns scaled after showing promise, with a reported 1.49x ROAS. Just one website in the paper's sample matched Opascope's performance on both conversion rate and implied revenue per visit — and it had only nine ChatGPT ad sessions. The authors report ChatGPT clicks cost about 2.4 times the search price while returning one-third the value, and a table shows a ChatGPT price 6.57 times what equal ROI would require. The Frankfurt School's release does not explain how its $0.52 and $3.40 figures were calculated; they appear nowhere in the paper itself.

What do agency case studies show?

  • StubGroup ran 30-day campaigns for a B2B compliance client. ChatGPT clicks cost $0.76 versus $10.46 on Google. Cost per lead was $92.21 versus $157.57. Conversion rates ran 0.8% versus 6.64%. Google's years of optimization mean its leads convert to customers at a higher rate so far; the client's long sales cycle leaves the cost-per-customer verdict open.
  • Workshop Digital spent $2,319 of its own money. ChatGPT CPC was $3.14 versus $6.76; CTR was 1.1% versus 5.2%. The agency got zero qualified leads from ChatGPT and two from Google, but did not disclose Google spend or dates, calling it "not a direct comparison."
  • Brafton spent $1,524.11 from June 17 to Aug. 28 for 263 clicks and one form submission. Its director of paid media, Zach Shah, said ChatGPT ads are "priced like search, but seem to behave like paid social."
  • OpenAI's own customer page cites Jotform, which reported cost per signup roughly 30% below paid search on internal July data. No costs or attribution method were disclosed, and sign-ups are not sales.

What did OpenAI say?

An OpenAI spokesperson said the study "captures an earlier period of ChatGPT Ads' development and extrapolates platform performance from a small sample." OpenAI also notes recent product launches including Target CPA and Target CPC. Its ads FAQ provides no performance benchmarks, and no OpenAI documentation discloses click prices across advertisers.

Measurement adds its own friction. Ads Manager tracks clicks and may include view-through conversions with 7-, 14-, or 30-day windows; the paper tracked ad-attributable sessions. The $0.52 figure only works if you assume clicks and sessions align. The paper's data also ended before OpenAI added conversion-optimized campaigns in June and bid adjustments this month. Since Nov. 9, new cost-target campaigns use Target CPA or Target CPC; legacy Max Bid and Max CPC retire Jan. 1, 2027. None of these options comes with demonstrated performance improvement.

For a paid search team, the defensible comparison starts with actual spend and attributed revenue on both channels, same period, same attribution settings — not another party's click prices.

via searchenginejournal.com (Source)

Filed under

  • chatgpt-ads
  • google-search-ads
  • advertising-roi
  • paid-media-measurement
  • attribution
Share this article:

More from Nathan Brooks

Nathan Brooks

Show full bio

Correspondent covering regulation, policy and their impact on marketing professionals running AI in the stack at Mart Signal.

154 articles

Bus out

‹ Previous article