ChatGPT Ads vs Google Ads and the Revenue Evidence

Google Search produced more tracked sales revenue per paid visit than ChatGPT ads in a recent academic comparison, a result that deserves examination by advertisers deciding where to allocate their next test budget. The commercial question concerns revenue earned by the advertiser. OpenAI's advertising income belongs to a separate discussion, and sales recorded after a click still require a margin calculation to establish profitability.

Frankfurt School published its account on November 3, 2026, following Maximilian Kaiser and Christian Schulze's October working paper. Google's advantage in that sample supports a cautious starting position for direct-response budgets. Application to a particular retailer requires its campaign evidence, especially when its products or customers differ from the research sample.

What the Study Actually Compared

Kaiser and Schulze's research covered 1,898 websites, with financial reporting available for a smaller subset comprising 102 advertisers. In the US paid-traffic comparison, tracked revenue averaged $0.89 per ChatGPT session versus $2.45 per Google Search session. The authors report a 64% shortfall for ChatGPT and conversion rates measuring 0.37% versus 1.17%.

Paid-channel comparisons were observational, with advertisers selecting their campaigns; the results establish a sample association with limited causal certainty. Data extended through August 22, 2026. The October working paper remains subject to academic review. Its paid-ad findings warrant separate treatment from its organic-referral analysis, which addresses a different research question.

For an advertiser, the relevant denominator deserves close inspection whenever a presentation compares revenue performance between platforms. Revenue per session measures recorded sales relative to website visits. Actual media expenditure enters a return-on-ad-spend calculation. Net commercial contribution requires further deductions, including product costs and fulfillment expenses. Confusion between those measures risks a budget decision built around the wrong economic target.

The Sixfold Price Claim Needs Its Assumptions

The study's price-parity estimate combines the revenue ratio with a $0.72 median Google CPC to estimate a ChatGPT CPC near $0.26 for equivalent returns. Its approximately sixfold pricing comparison uses an externally disclosed $1.72 ChatGPT CPC from one account. Advertiser-wide observed ChatGPT costs would provide a firmer basis for a general price conclusion.

Conditional arithmetic can help a buyer set an initial bid ceiling, provided the assumptions remain visible beside the estimate. Comparable profit margins are necessary when a revenue ratio becomes a proxy for commercial value. Different products can carry substantially different margins. Revenue from a heavily discounted product may support less acquisition expenditure than a smaller sale carrying a healthier contribution.

Consider the decision facing a retailer whose Google campaigns already produce acceptable returns at their current budget. Expenditure transferred into ChatGPT creates an opportunity cost equal to the contribution that the displaced Google activity would have produced. Additional testing funds face a different comparison: the expected learning value and potential future contribution must justify the cash spent. Budget managers should specify which decision they are making when they approve the experiment.

November Reports Need Current Measurement Settings

OpenAI announced additional measurement partnerships on October 5, after the paper's data period, and described geo-based incrementality work as early-stage research. Its current Ads Manager documentation offers configurable click-through windows and optional one-day view-through reporting. Reported totals can include purchases credited after an impression and purchases following a click.

OpenAI's announcement includes advertiser examples with favorable results, published by the platform and attributed to measurement partners. Published cases concern particular campaigns with their own methods. Comparison with the working paper requires the original study design and commercial population for each claim. Averages collected using different attribution rules answer different questions, even when their dashboard labels look similar.

Advertisers reviewing both platforms should record the selected attribution window beside the revenue figure used in the comparison. Google Ads documentation explains that attribution choices affect conversion credit and automated bidding. Identical report dates offer limited comparability when one report uses impression-based credit and the other uses click-based credit. Separate columns for those categories make the comparison easier to inspect.

Establish a Commercial Test That Can Be Evaluated

Successful experimental planning begins with an explicit purchasing outcome and the financial threshold that would justify further expenditure for the business. Retail teams can nominate completed orders as the primary outcome. Subscription businesses need a separately specified definition for paid activation, with the assessment period long enough to observe the relevant billing event. Trial registration and a paying customer merit distinct values in the financial review.

Next, assign a finite test budget and retain a comparable reference period for the existing Google activity. Product availability should remain comparable during the assessment. Any major promotion deserves its own annotation, since discounts affect both conversion probability and sales value. Traffic composition also needs scrutiny: established customers may behave differently from people encountering the retailer for the first time.

The finance team should agree on how refunds will be included in the evaluation, while campaign owners establish the minimum observation period for delayed purchases. Recent sales can overstate retained revenue when cancellations arrive after the initial reporting cutoff. Delayed purchases create the opposite problem for an early assessment. The reporting calendar should accommodate the business's actual purchase cycle and return policy.

Follow the Order Past the Advertising Dashboard

Order-level reconciliation provides a useful safeguard when two platforms credit the same purchase during overlapping attribution windows. Compare attributed sales with the commerce system's accepted orders, then document how duplicates enter the channel analysis. Merchant records provide the revenue total against which platform claims can be reconciled. Consent requirements and applicable privacy rules govern the use of customer data in that work.

For lead-generation advertisers, a completed inquiry requires an additional link to the eventual sales outcome to support a revenue judgment. CRM records can connect an inquiry to an accepted proposal. Qualified opportunities may justify a provisional value during a long sales cycle, provided that estimate uses a documented historical conversion rate. Closed sales should replace provisional values when the evidence becomes available.

OPERENDIA's Google Ads Management service addresses conversion tracking alongside account review, a relevant foundation for evaluating an additional paid channel. Campaign reporting becomes commercially useful when the people buying media can inspect the same sales evidence as the people responsible for revenue. The assessment should identify the decisions supported by observed orders and the decisions still dependent on estimates.

Reported averages can conceal an expensive mismatch between a campaign's targeting and the merchandise available on its destination page. Advertisers should examine that explanation alongside media price when a test yields weak sales. Poor landing-page relevance would require a different response from a campaign whose purchase rate is acceptable but whose clicks cost too much. The diagnosis determines which variable deserves the next test.

Decide What Additional Budget Must Prove

Businesses considering more ChatGPT advertising should set their expansion criteria using their contribution requirements and the uncertainty remaining after the first test. Positive attributed revenue provides an initial signal. Sustained spending requires evidence that the resulting sales justify acquisition costs, with an allowance for operational expenses. Volume should enter the decision only after the economics have been examined.

Incrementality testing addresses the question of additional business attributable to advertising, including purchases that customers would have made independently of the campaign. Where test design and available volume permit a credible comparison, a holdout can help estimate that additional contribution. Smaller campaigns may yield estimates with wide uncertainty. Expenditure limits should remain proportionate to the evidence available for the next decision.

Google Search currently has the stronger observed revenue-per-session result in the cited comparison, and the reported ChatGPT cost benchmark makes price scrutiny especially relevant. Future allocation still requires current campaign data evaluated against a consistent commercial objective. For the next budget review, request retained sales revenue and acquisition expenditure for both channels, along with the attribution settings that produced these figures.