Krystal Scanlon · Nov 5, 2026
Digiday: Figures show ChatGPT ads growing, but conversion remains a problem
Markdown · JSON · Original source
’s Krystal Scanlon argues that ChatGPT advertising is expanding, while evidence that it delivers results for advertisers remains weaker than for paid Google Search. Spending is spreading beyond shopping, and Adthena tracked a growing number of brands advertising on the platform across the markets it measured. But those signs of demand do not show that the ads are converting well. A working paper under review, based on ecommerce website data, reported lower revenue per session and conversion rates for paid ChatGPT visits than for paid Google Search visits. Scanlon also reports that New Street Research cautions its estimate of CPC may be too high because monetization may be weaker than it modelled. Accuracast CEO Farhad Divecha says partner integrations have not necessarily resolved tracking issues, including discrepancies he attributes partly to attribution problems. The article’s argument is that advertiser growth and ad opportunities are not enough to establish value: advertisers need evidence that ads deliver results, and measurement problems make that harder to assess. The reported comparison is limited to ecommerce data. How much do tracking and attribution problems affect the reported performance gap?
Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.
Original source
· Article
Source details
Original source date: Nov 5, 2026 (2026-11-05T05:01:00+00:00).
Date basis: Date stated by the publisher.
Added to Escape Meridian: 2026-11-05T13:53:39.873Z. First observed: 2026-11-05T05:03:26.593Z. Retrieved: 2026-11-05T13:52:31.500Z.
Referenced by
Recorded citations and public file attachments. A reference alone does not establish agreement or influence.
No public references recorded here yet.
Limits
The reported performance comparison comes from a working paper under review and is based on data from ecommerce websites; the excerpt does not establish whether the findings generalize to other types of advertising or advertisers. The article reports Adthena data only through September 20 for September and describes the full-month as an estimate, not a completed count. The article reports that tracking and attribution problems persist but does not quantify their effect on the reported performance comparison.
Open question
How much do tracking and attribution problems affect the reported performance gap between paid ChatGPT visits and paid Google Search visits?
Source-stated
Scanlon reports that the share of ChatGPT ad spending attributed to shopping declined as financial services, software, and travel and tourism grew their shares between April and September.
shopping accounted for 38% of U.S. ad spend on the platform in April, but that share fell to 20% by September.
Source-stated
Adthena tracked a rapid increase in distinct brands running ChatGPT ads across the U.S., Australia, the U.K. and the EU from June through August; the September figure was a projection based on data through September 20.
That number rose to 3,741 in July and 6,333 in August.
Source-stated
A working paper under review, based on ecommerce website data, reported lower revenue per session and conversion rates for paid ChatGPT visits than for paid Google Search visits.
paid ChatGPT visits generate $0.89 in revenue per session for U.S. advertisers, compared with $2.45 for paid Google Search visits.
Source-stated
New Street Research cautioned that monetization may be weaker than it had modelled, so its estimated CPC could be too high.
suggesting its estimated CPC of around $5 could be too high.
Source-stated
Accuracast CEO Farhad Divecha said partner integrations had not necessarily resolved tracking problems and described continuing discrepancies.
“We continue to see discrepancies there,” he said. “Some are teething problems because it’s a new system that has to be set up on advertiser systems, but some problems are down to attribution being messed up.”
Source and contributor context
Krystal Scanlon · Article author
These are credits stated by the source; they do not establish a profile match or authorship of every passage.
Source media
Media declared by the original publisher; availability checked separately from the editorial reading.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
Digiday
Digiday is a trade publication covering digital media, marketing and advertising businesses.
Figure
Figure develops humanoid robots and publishes its own tests of how they perform physical tasks.
Editorial provenance
Version 2. Publication ID: fc4afea6-6272-4b8e-8ed8-22e2c7134240.
Prepared by gpt-6-luna; grounding review by gpt-6-luna.
Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-05T13:53:39.873Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: useful_perspective.
Observation: 3365559ea07793cc76353ff335c51efa59ecc627221e88b9ca47c1d6d0c8f649. Source text hash: 10fcafb26b4bfd82fbd6ea023542d2f2d5f1648fe8ccba2b7589ef372f145b6d.
Correction reason: The revised summary makes Scanlon’s central tension clearer: advertiser growth does not establish performance, and unresolved measurement issues complicate assessment.
Revision and source-check history
Version 2 · 2026-11-05T13:53:39.873Z · The revised summary makes Scanlon’s central tension clearer: advertiser growth does not establish performance, and unresolved measurement issues complicate assessment.
Digiday: Figures show ChatGPT ads growing, but conversion remains a problem
’s Krystal Scanlon argues that ChatGPT advertising is expanding, while evidence that it delivers results for advertisers remains weaker than for paid Google Search. Spending is spreading beyond shopping, and Adthena tracked a growing number of brands advertising on the platform across the markets it measured. But those signs of demand do not show that the ads are converting well. A working paper under review, based on ecommerce website data, reported lower revenue per session and conversion rates for paid ChatGPT visits than for paid Google Search visits. Scanlon also reports that New Street Research cautions its estimate of CPC may be too high because monetization may be weaker than it modelled. Accuracast CEO Farhad Divecha says partner integrations have not necessarily resolved tracking issues, including discrepancies he attributes partly to attribution problems. The article’s argument is that advertiser growth and ad opportunities are not enough to establish value: advertisers need evidence that ads deliver results, and measurement problems make that harder to assess. The reported comparison is limited to ecommerce data. How much do tracking and attribution problems affect the reported performance gap?
Original source
· Article
Source details
Original source date: Nov 5, 2026 (2026-11-05T05:01:00+00:00).
Date basis: Date stated by the publisher.
Added to Escape Meridian: 2026-11-05T13:53:39.873Z. First observed: 2026-11-05T05:03:26.593Z. Retrieved: 2026-11-05T13:52:31.500Z.
Limits
The reported performance comparison comes from a working paper under review and is based on data from ecommerce websites; the excerpt does not establish whether the findings generalize to other types of advertising or advertisers. The article reports Adthena data only through September 20 for September and describes the full-month as an estimate, not a completed count. The article reports that tracking and attribution problems persist but does not quantify their effect on the reported performance comparison.
Open question
How much do tracking and attribution problems affect the reported performance gap between paid ChatGPT visits and paid Google Search visits?
Source-stated
Scanlon reports that the share of ChatGPT ad spending attributed to shopping declined as financial services, software, and travel and tourism grew their shares between April and September.
shopping accounted for 38% of U.S. ad spend on the platform in April, but that share fell to 20% by September.
Source-stated
Adthena tracked a rapid increase in distinct brands running ChatGPT ads across the U.S., Australia, the U.K. and the EU from June through August; the September figure was a projection based on data through September 20.
That number rose to 3,741 in July and 6,333 in August.
Source-stated
A working paper under review, based on ecommerce website data, reported lower revenue per session and conversion rates for paid ChatGPT visits than for paid Google Search visits.
paid ChatGPT visits generate $0.89 in revenue per session for U.S. advertisers, compared with $2.45 for paid Google Search visits.
Source-stated
New Street Research cautioned that monetization may be weaker than it had modelled, so its estimated CPC could be too high.
suggesting its estimated CPC of around $5 could be too high.
Source-stated
Accuracast CEO Farhad Divecha said partner integrations had not necessarily resolved tracking problems and described continuing discrepancies.
“We continue to see discrepancies there,” he said. “Some are teething problems because it’s a new system that has to be set up on advertiser systems, but some problems are down to attribution being messed up.”
Source and contributor context
Krystal Scanlon · Article author
These are credits stated by the source; they do not establish a profile match or authorship of every passage.
Source media
Media declared by the original publisher; availability checked separately from the editorial reading.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
Digiday
Digiday is a trade publication covering digital media, marketing and advertising businesses.
Figure
Figure develops humanoid robots and publishes its own tests of how they perform physical tasks.
Editorial provenance
Version 2. Publication ID: fc4afea6-6272-4b8e-8ed8-22e2c7134240.
Prepared by gpt-6-luna; grounding review by gpt-6-luna.
Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-05T13:53:39.873Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: useful_perspective.
Observation: 3365559ea07793cc76353ff335c51efa59ecc627221e88b9ca47c1d6d0c8f649. Source text hash: 10fcafb26b4bfd82fbd6ea023542d2f2d5f1648fe8ccba2b7589ef372f145b6d.
Correction reason: The revised summary makes Scanlon’s central tension clearer: advertiser growth does not establish performance, and unresolved measurement issues complicate assessment.
Version 1 · 2026-11-05T05:38:37.961Z
Digiday: Six thousand brands later, ChatGPT ads still have a conversion problem
’s Krystal Scanlon argues that ChatGPT advertising is expanding faster than evidence that it can deliver results comparable to established performance channels. Advertiser spending is spreading beyond shopping, and Adthena tracked a sharp rise in brands running ads across the markets it measured. But growth in advertisers and ad opportunities does not establish that those ads convert well. A working paper under review, using data from ecommerce websites, found lower revenue per session and conversion rates for paid ChatGPT visits than for paid Google Search visits. Scanlon also reports that New Street Research’s model may overestimate CPC: the firm cautions that monetization may be weaker than it had originally modelled. Accuracast CEO Farhad Divecha says partner integrations have not necessarily resolved tracking issues, including discrepancies he attributes partly to attribution problems. The article’s central tension is between a growing advertising business and unresolved evidence about advertiser returns. The reported performance comparison is limited to the study’s ecommerce data, while tracking problems complicate measurement. Whether ChatGPT ads can demonstrate comparable value as the platform expands remains an open question.
Original source
· Article
Source details
Original source date: Nov 5, 2026 (2026-11-05T05:01:00+00:00).
Date basis: Date stated by the publisher.
Added to Escape Meridian: 2026-11-05T05:38:37.961Z. First observed: 2026-11-05T05:03:26.593Z. Retrieved: 2026-11-05T05:37:24.443Z.
Limits
The reported performance comparison comes from a working paper under review and is based on data from ecommerce websites; the excerpt does not establish whether the findings generalize to other types of advertising or advertisers. The article reports Adthena data only through September 20 for September and describes the full-month as an estimate, not a completed count. The article reports that tracking and attribution problems persist but does not quantify their effect on the reported performance comparison.
Open question
How much do tracking and attribution problems affect the reported performance gap between paid ChatGPT visits and paid Google Search visits?
Source-stated
Scanlon reports that the share of ChatGPT ad spending attributed to shopping declined as financial services, software, and travel and tourism grew their shares between April and September.
shopping accounted for 38% of U.S. ad spend on the platform in April, but that share fell to 20% by September.
Source-stated
Adthena tracked a rapid increase in distinct brands running ChatGPT ads across the U.S., Australia, the U.K. and the EU from June through August; the September figure was a projection based on data through September 20.
That number rose to 3,741 in July and 6,333 in August.
Source-stated
A working paper under review, based on ecommerce website data, reported lower revenue per session and conversion rates for paid ChatGPT visits than for paid Google Search visits.
paid ChatGPT visits generate $0.89 in revenue per session for U.S. advertisers, compared with $2.45 for paid Google Search visits.
Source-stated
New Street Research cautioned that monetization may be weaker than it had modelled, so its estimated CPC could be too high.
suggesting its estimated CPC of around $5 could be too high.
Source-stated
Accuracast CEO Farhad Divecha said partner integrations had not necessarily resolved tracking problems and described continuing discrepancies.
“We continue to see discrepancies there,” he said. “Some are teething problems because it’s a new system that has to be set up on advertiser systems, but some problems are down to attribution being messed up.”
Source and contributor context
Krystal Scanlon · Article author
These are credits stated by the source; they do not establish a profile match or authorship of every passage.
Source media
Media declared by the original publisher; availability checked separately from the editorial reading.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
Digiday
Digiday is a trade publication covering digital media, marketing and advertising businesses.
Figure
Figure develops humanoid robots and publishes its own tests of how they perform physical tasks.
Editorial provenance
Version 1. Publication ID: c2b9a852-300e-44da-8e9e-8625bb23099f.
Prepared by gpt-6-luna; grounding review by gpt-6-luna.
Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-05T05:38:37.961Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: useful_perspective.
Observation: 0ac443cf2220af0cd057c6688ece4602d8672425f69454f5c9b35e9c28562fc0. Source text hash: efdb8567183b959e532d6e3996581eae1450bf4712680bd0e22d9029a1b18771.
uncertain · 2026-11-05T05:38:37.961Z · anthropic/claude-opus-5.5
Publication: c2b9a852-300e-44da-8e9e-8625bb23099f
attribution — pass: The reading correctly attributes the article and framing to Krystal Scanlon of Digiday. It also attributes the data to Sensor Tower, Adthena, New Street Research and the Kaiser/Schulze working paper, and the quote to Farhad Divecha. The source calls Divecha Accuracast's "group CEO"; shortening this to "CEO" is minor and does not change the attribution.
according to Accuracast’s group CEO Farhad Divecha
summary — uncertain: The reading faithfully captures the article's central tension: advertising growth is outrunning proof of advertiser returns. However, "CPC" appears in the summary without explanation, and the article's own gloss (pricing per click versus per impression) is omitted. Smallest repair: spell out cost per click (CPC) or add a term-library tooltip. "Conversion" may also merit a tooltip. The title and question are fair, and the question follows from the reported tracking problems.
given CPC ads typically generate more per-ad event than impression-based pricing
date — pass: The displayed original publication date, 2026-11-05, matches the article byline date.
By Krystal Scanlon • November 5, 2026
support — uncertain: Most claims are directly supported, and the limitations accurately note the September projection and the working paper's ecommerce scope. Two scope conditions are dropped, though. Claim 1 describes "ChatGPT ad spending" generally, but the Sensor Tower shares are U.S. ad spend. The summary's revenue-per-session comparison also omits that the figures are for U.S. advertisers. Smallest repair: add "U.S." to both passages.
shopping accounted for 38% of U.S. ad spend on the platform in April
media — pass: The selected media is the verified direct link to the original Digiday article, which is the most direct way for readers to reach the source.
uncertain · 2026-11-05T13:53:39.873Z · anthropic/claude-opus-5.5
Publication: fc4afea6-6272-4b8e-8ed8-22e2c7134240
attribution — pass: The byline credits Krystal Scanlon at Digiday, and the draft attributes each data point to its named source: Sensor Tower spend shares, Adthena brand counts, New Street Research, the Kaiser and Schulze working paper, and Divecha. Divecha's title is 'group CEO' of Accuracast, while the draft says 'CEO'. The difference is minor and not misleading.
according to Accuracast’s group CEO Farhad Divecha
summary — uncertain: CPC is central to the New Street point but is never explained. A tooltip or plain gloss ('cost per click') is needed, and 'conversion' and 'attribution' would benefit from tooltips too. The draft also omits the article's framing that OpenAI only began selling ads this year (the pilot launched in February), so lagging ad performance is expected. That condition is part of how the source positions the gap. Smallest repair: add that OpenAI started selling ChatGPT ads this year and gloss CPC. The summary's closing sentence also repeats the separate question.
Nobody should be surprised by that. OpenAI only started selling ads this year
date — pass: The originalPublished value, 2026-11-05T05:01 UTC, matches the article byline date of November 5, 2026 and the feed entry. No earlier publication is indicated.
By Krystal Scanlon • November 5, 2026
support — uncertain: The claims are source-stated, and the excerpts are exact. Scope is narrowed in places, however. The shopping and spend-share shift is Sensor Tower data on U.S. ad spend, yet the draft states it generally as 'ChatGPT ad spending'. The working-paper figures are also for U.S. advertisers. Smallest repair: add 'U.S.' to the spend-share claim and to the performance comparison. The limitations correctly note that the September Adthena figure is a projection and that the paper is under review.
According to data from Sensor Tower, shopping accounted for 38% of U.S. ad spend on the platform in April
media — pass: The selected media is the verified direct link to the original Digiday article. It is the most direct way for readers to access the source.
Challenge or correct this reading
Send a correction or an opposing source with this reading’s URL. Disagreement and uncertainty can remain unresolved.
Digiday is a trade publication covering digital media, marketing and advertising businesses.
Explore DigidayDigiday is a trade publication covering digital media, marketing and advertising businesses.
Explore DigidayFigure develops humanoid robots and publishes its own tests of how they perform physical tasks.
Explore FigureDigiday is a trade publication covering digital media, marketing and advertising businesses.
Explore DigidayDigiday is a trade publication covering digital media, marketing and advertising businesses.
Explore DigidayFigure develops humanoid robots and publishes its own tests of how they perform physical tasks.
Explore FigureDigiday is a trade publication covering digital media, marketing and advertising businesses.
Explore DigidayDigiday is a trade publication covering digital media, marketing and advertising businesses.
Explore DigidayFigure develops humanoid robots and publishes its own tests of how they perform physical tasks.
Explore Figure