Three accounts published on November 2 looked at the same budget line from different ends. Digiday asked media buyers how much of their clients’ Google search spend is under review. Search Engine Journal reported on a working paper that prices a ChatGPT ad visit against a Google search ad visit. MarTech.org argued that 2027 media plans start from the wrong unit when they start from channels. Read together, they describe a search budget that buyers want to move and an evidence base that does not yet show where it should go.
What the three accounts report
Digiday: buyers put Google under review
Digiday’s report draws on conversations with 18 media buyers and finds clients putting Google through the kind of scrutiny usually saved for newer channels. Its account of the cause is mechanical. AI Overviews answer more queries on the results page, which leaves fewer prime ad slots, and Google’s automated bidding pays up for the ones that remain while Q4 seasonal advertisers push cost per click higher.
The buyers put numbers on it. Lindsay Marty, founder of Above the Bar Marketing, told Digiday: “The amount being questioned is generally around 10% to 20% of the Google budget when that conversation comes up.” Bryce Hamlin, vp of data intelligence at Power Digital, described marketers asking whether a portion of the budget is better spent elsewhere, “to the tune of 10% to 30% of the Google search budget getting reallocated.” Accuracast group CEO Farhad Divecha put his clients’ figure at about 10% for now, with room to reach 20% to 30% around Black Friday and Cyber Monday. Digiday also cites IAB and PwC figures that search took $114.2 billion of U.S. digital ad spending last year, 38.8% of the total.
Search Engine Journal: the tested alternative earns less per visit
Search Engine Journal’s report covers a working paper by Maximilian Kaiser of the Vienna University of Economics and Business and Christian Schulze of the Frankfurt School of Finance and Management. Using Google Analytics e-commerce data from online shops, it follows traffic and sales through OpenAI’s country-by-country ad launch. In U.S. traffic, a visit from a ChatGPT ad brought in $0.89 in revenue against $2.45 for a visit from a paid Google search ad. ChatGPT ad visits converted at 0.37% against 1.17% for paid search.
The spread across advertisers is the striking part. Of 87 advertisers that bought both channels in U.S. traffic, 84 earned less per ChatGPT ad visit than per paid search visit, and 73 recorded no sales from ChatGPT ads at all. One advertiser accounted for 66% of all purchases made through paid ChatGPT. The authors use paid search as the benchmark for a parity price and arrive at about $0.26 per ChatGPT click, against a median paid search click cost of $0.72 in their data. SEJ notes that OpenAI’s own guidance recommends starting cost-per-click campaigns at a maximum bid of $3 to $5.
The paper’s own summary page lists it as version 1, dated October 2026 and under review, built on 1,898 websites over 60 weeks. It states that paid ChatGPT visits offer less than a sixth of the value for money of paid Google Search, and that organic ChatGPT referrals fell by about 6% to 8% after ads arrived while the conversion rate of those visits held. SEJ adds the paper’s limits: it sees the last click only, without prompts, impressions or subscription tiers, and advertisers chose to buy ChatGPT ads, so the comparison describes their results without proving the ads caused them.
MarTech.org: plan by audience before channel
MarTech.org’s column addresses the same Q4 planning cycle from the measurement side. It argues that channel budgets and reporting should support an audience strategy split into new, occasional and loyal customers, each with its own KPIs and media mix. Its example is a search campaign reporting a 6:1 return on ad spend. The column says that figure shows what search received credit for and leaves open what it caused, and that the next dollar will not necessarily return the same if other media helped create the demand. It recommends attribution, modelling and incrementality testing to establish how platforms work individually and together for each segment.
What Google’s own numbers say
Alphabet’s third-quarter release, filed with the SEC on October 28, reports Google Search and other revenue of $63,350 million, up from $56,567 million a year earlier, a rise of 12%. Digiday describes that as Alphabet’s slowest pace this year. Sundar Pichai, CEO of Google and Alphabet, said in the release: “Our AI investments are driving overall Search query growth, including commercial queries.” The release says year-over-year query growth was the highest in five years. Digiday’s buyers make a similar point from their side, reporting that impressions and clicks climbed along with cost per click.
Where the accounts disagree
The accounts split on what a buyer should do with the doubt. Digiday’s buyers talk in reallocation shares, 10% to 30% of the line, with Meta and Local Services Ads named as destinations if performance weakens. The working paper tests the newest destination, ads inside an AI assistant, and finds that a visit from it returned about a third of a paid search visit’s revenue in U.S. traffic. MarTech.org would reject the framing of both, because a share moved between channels says nothing about which customers it reaches or whether it changes what they buy.
They also differ on what counts as evidence. Digiday’s sources cite media mix models and geo-tests that credit upper-funnel media more and search less. The working paper uses last-click revenue per visit, the measure MarTech.org warns overstates a channel that harvests demand others created. A buyer could hold all three views at once and still lack a number to act on.
On one point all three agree. The value of a Google search click is no longer taken as given, and the Q4 planning window is when that question gets asked of real budgets.
What the coverage adds up to
In our reading, the week’s coverage shows pressure on search budgets running ahead of the evidence for moving them. Buyers are naming shares to move. The one rigorous published comparison of the newest alternative found it earning less per visit, with sales concentrated in a single advertiser. Google reports more queries and more commercial queries, and higher prices for the clicks around them. The case for moving money therefore rests on incrementality tests that each advertiser has to run for itself, because the shared data so far favours search on a per-visit basis.
Our earlier reports on Google’s AI Mode ads expansion and on how ChatGPT ads are measured cover the inventory and attribution questions behind those numbers.
What it means for the marketing leader
Treat the 10% to 30% figures as the size of the test, and fund the test before the move. A geo-holdout on the share of search you are considering cutting will show whether the conversions it gets credit for disappear when it is switched off. If you test ChatGPT ads, compare revenue per visit against your own paid search, and check how concentrated any sales are before scaling, since the working paper found 73 of 87 advertisers with no ChatGPT ad sales. Ask your search agency to report impressions, clicks and cost per click together for Q4, so a price rise and a volume rise are visible separately. Our opinion piece on market-level reporting for AI answer ads lists the reporting to request before budget follows.
