ChatGPT Ads Are Here. What Should Advertisers Measure?

OpenAI has introduced new advertising formats and expanded its measurement tools, while Adobe has added support for managing ChatGPT campaigns. Marketers can now compare the platform with established advertising channels, although determining its contribution to sales remains difficult.

OpenAI is expanding advertising in ChatGPT with a new visual format and additional measurement tools, as it develops a commercial business around a platform used by 1.2 billion people each week.

Announced on 5 October 2026, the changes include partnerships with attribution specialists AppsFlyer, Northbeam, Triple Whale and DV Rockerbox. Adobe has also introduced beta support for ChatGPT Ads in its Advertising Search, Social & Commerce platform, allowing marketers to manage campaigns and monitor performance alongside their existing advertising activities.

For advertisers, the appeal lies in reaching people while they research products, weigh alternatives and consider purchases. Measuring the value of those encounters presents a familiar challenge. Clicks and conversions can show what happened after someone saw an advertisement, but they offer less certainty about whether the advertising persuaded them to buy something they would otherwise have passed over.

OpenAI’s new tools should help marketers examine that relationship, although ChatGPT advertising has a much shorter performance record than established search and social platforms.

How Advertising in ChatGPT Works

Conventional search advertising generally matches advertisements to search terms. ChatGPT can also consider the subject and apparent intent of a conversation when selecting which advertisements to display.

A user researching equipment for a home office, for example, might see an advertisement for a relevant product or service. OpenAI assesses the conversational context alongside an advertiser’s creative material, landing page and targeting preferences.

Advertisements appear separately from ChatGPT’s responses and carry clear labels. OpenAI says advertisers cannot influence the answers its models produce or access users’ private conversations.

The platform began testing advertisements in the United States in February 2026, initially showing them to eligible users of its Free and Go plans. Paid subscriptions, including Plus and Pro, remain free of advertising.

OpenAI’s latest format uses images to present products and services, giving advertisers more scope to show what they sell. The company plans to test these placements during image generation in October, beginning with a selected group of American advertisers.

Visual advertising could be particularly attractive to retailers selling products whose appearance influences purchasing decisions. For now, marketers need to separate the advertising tools already available from formats that OpenAI has announced but has yet to roll out more widely.

Measuring Sales, Not Just Clicks

ChatGPT Ads Manager provides familiar campaign figures, including impressions, clicks, spending and click-through rates.

Advertisers can also use OpenAI’s Pixel and Conversions API to record actions such as purchases, registrations and enquiries. The October update adds integration options through Hightouch, Tealium and LiveRamp, while attribution partners can help businesses bring ChatGPT campaign data into their existing reports.

Adobe’s integration gives marketing teams another way to manage the channel. Advertisers using its Advertising Search, Social & Commerce platform can create campaigns, organise ad groups, manage creative material and monitor impressions, clicks and costs without handling every task separately in OpenAI’s system.

The integration has limitations, however. Adobe does not currently display conversions tracked by ChatGPT Ads, so marketers must consult OpenAI’s Ads Manager for those figures. Several of Adobe’s established optimisation and analytics features are also unavailable for ChatGPT campaigns.

Businesses can therefore manage some of their advertising activity in one place, but assessing performance still requires information from different systems.

What the Early Results Tell Advertisers

OpenAI has released initial campaign results through its measurement partners, offering some indication of how advertisers are using the platform.

DV Rockerbox reported that WeightWatchers achieved a cost per acquisition through ChatGPT Ads that was 15.3% lower than its blended paid-search benchmark.

Triple Whale found that 93% of visitors reaching Portland Leather through ChatGPT advertisements were new to the retailer’s website. WorkMagic reported that 67% of incremental purchases identified in a study of wellness brand Dose came from new customers.

The figures suggest that ChatGPT advertising can introduce brands to people they may not have reached through other channels. Whether it does so consistently will depend on the advertiser, the product and the campaign.

Nor do the findings all measure the same thing. A first-time website visitor may already know the brand, and an attributed purchase does not necessarily mean the advertisement persuaded the customer to buy.

That matters when businesses are deciding whether to add ChatGPT to their advertising budgets. A campaign that produces apparently inexpensive conversions may still deliver little additional revenue if most customers would have bought the product anyway.

The Problem with Attribution

Customers do not always purchase immediately after seeing an advertisement, and conversations with an AI assistant may involve several stages of research before anyone visits a retailer’s website.

Someone comparing financial software, for example, might encounter an advertisement in ChatGPT, continue exploring other options and eventually reach the advertiser through a search engine or another website.

A reporting system that credits only the final click may overlook ChatGPT’s earlier involvement.

OpenAI has introduced view-through conversion reporting to capture qualifying customer actions that follow an advertising impression, even when the person does not click the advertisement.

In an early analysis of global campaigns optimised for actions further along the purchasing journey, OpenAI found that 52.7% of eligible one-day view-through conversions occurred within an hour of the advertising impression.

The finding shows how quickly some measurable customer activity follows exposure to an advertisement. It cannot establish whether those customers would have acted differently without seeing it.

OpenAI is also working with Haus, Measured and WorkMagic on geographic experiments designed to estimate additional sales. These studies compare the results in markets exposed to advertising with those in comparable control markets.

WorkMagic, for example, estimated that Dose generated 2.3 times as many incremental orders as its last-click reporting recorded.

Such experiments can provide stronger evidence of advertising effectiveness because they attempt to measure what would have happened without the campaign. Their reliability nevertheless depends on how the groups are selected, how long the experiment runs and whether the results are statistically convincing.

Advertisers should examine those details before treating early case studies as a reliable indication of the returns they can expect.

Brand Safety Needs Its Own Assessment

ChatGPT’s conversational format also creates questions about where advertisements appear.

A product may be relevant to the subject of a conversation without being suitable for the particular circumstances in which that subject arises.

OpenAI says it screens conversations for sensitive or unsuitable contexts and allows eligible advertisers to exclude subjects through Negative Phrases.

The company is also developing independent assessments with DoubleVerify and Integral Ad Science, which will examine its advertising safeguards without giving the external companies access to private conversations.

Businesses with strict requirements governing advertising placements should review these protections before launching campaigns, particularly in sectors where promotional activity faces additional restrictions.

They should also confirm which controls are available to their accounts rather than assuming that every announced safeguard is already in place.

Five Questions Advertisers Should Ask Before Increasing Their Budgets

Marketing teams considering ChatGPT Ads should begin with a specific sales or customer-acquisition target and establish how they intend to measure progress against it.

1. Are the advertisements bringing in new customers?

Compare the proportion of new buyers with results from existing campaigns. A first-time website visitor is not necessarily a first-time customer, so businesses should use their own sales records where possible.

2. How much does each customer cost?

Measure the cost of acquiring a customer or generating a qualified lead against comparable search and social campaigns. Use the same conversion definitions and reporting periods, otherwise the comparison may be misleading.

3. Is conversion tracking reliable?

Check that OpenAI’s Pixel or Conversions API records the intended actions accurately. Duplicate events, missing purchases or inconsistent attribution settings can distort reported performance.

4. Is the campaign producing additional sales?

Where budgets and conversion volumes are sufficient, controlled experiments can help determine whether ChatGPT advertising increases total sales or merely receives credit for purchases that would have occurred anyway.

5. Are the advertisements appearing in suitable contexts?

Review the available placement controls, check that advertisements and landing pages describe the product accurately, and identify any subjects the brand should avoid.

Campaign performance also needs to be monitored as spending increases. A small test may reach the most receptive customers first, producing results that become harder to maintain at a larger scale.

Setting a budget, testing over a defined period and agreeing in advance which figures will determine the next investment decision should make the initial assessment more reliable.

What ChatGPT Ads Mean for Marketing Budgets

ChatGPT gives advertisers access to customers during conversations in which they may be researching a purchase, comparing products or seeking advice. OpenAI’s measurement partnerships and Adobe’s campaign-management support make those opportunities easier to test alongside existing advertising channels.

The early case studies offer encouraging results, although the platform has yet to establish a comparable record across different industries, budgets and campaign objectives.

Marketers considering ChatGPT Ads should therefore treat the channel as an experiment with a measurable commercial objective. Its conversational format may change how customers encounter advertising, but businesses still need to establish how many additional customers it brings in, what those customers spend and whether the returns justify a larger budget.