OpenAI adds visual ads and expanded measurement to ChatGPT

The ads use images to show product inspiration, how a product is used or experiences it could make possible.[2] OpenAI says the format is meant to help people discover brands, understand what makes an offering distinctive and decide what to buy.[2] The initial test will run during image generation in ChatGPT, with an initial group of advertisers in the United States beginning later in October.[2] OpenAI says ads will be clearly labeled and kept separate from the image being created, and that advertising does not influence ChatGPT’s answers.[2]

OpenAI says its expanded measurement tools are designed to work with both in-house and partner-led solutions.[2] Integrations with Hightouch, Tealium and LiveRamp let advertisers send conversion data from their existing systems to ChatGPT Ads.[2] The pixel and Conversions API measure actions including purchases, leads, installs and sign-ups, while Event Quality Score is intended to help advertisers assess and improve their conversion signals.[1] Advertisers can select a click-through attribution window and add a one-day view-through window for conversions after an impression without a click; those settings change reporting, not campaign optimization.[1]

OpenAI named measurement partners including AppsFlyer, Triple Whale, Adjust, DV Rockerbox, Northbeam, Branch, Singular, Kochava, Airbridge and Tenjin, alongside full-funnel and advanced measurement partners Fospha, Measured and INCRMNTAL.[2] It is also working with Haus, Measured and WorkMagic on early-stage geo-based experiments to measure causal impact.[2] DV Rockerbox reported that WeightWatchers’ attributed cost per acquisition on ChatGPT Ads was 15.3% below its blended paid-search benchmark.[2] WorkMagic reported statistically significant lift for Dose, with 67% of incremental purchases coming from new customers; its geo-lift study estimated 2.3 times more incremental orders than last-click attribution captured.[2][1]

Businesses can sign up for ads today.[2] OpenAI says its placement guardrails assess whether a conversation is appropriate for advertising and aim to exclude emotionally vulnerable, sensitive or otherwise unsuitable contexts.[2] The company says it uses automated review, human oversight and ongoing monitoring to identify policy violations and improve safeguards.[2] Qualifying advertisers with narrow placement requirements tied to their own policies can use Negative Phrases, and independent partners can assess brand-safety standards in a controlled environment without access to private user conversations.[2] IAS will independently evaluate how the safeguards perform across a broad and evolving range of scenarios.[2]