Before you raise the budget on a holiday ChatGPT Ads campaign, trace one completed order from checkout to Ads Manager. Can you confirm that the right data source received it, that the purchase was counted once, and that the campaign was set up to give an eligible ad interaction credit?
That is the measurement takeaway from OpenAI’s 4 November ChatGPT Ads holiday webinar. OpenAI’s public holiday guide reports that U.S. product-search activity in ChatGPT rose 94% on Black Friday 2025 compared with matched weekdays in the previous four weeks. The figure describes shopping research in one market. Your own campaign needs a baseline for sales or qualified enquiries before the seasonal rush. OpenAI’s ChatGPT Ads holiday guide.
What does Event Quality Score tell you?
Event Quality Score (EQS) assesses how useful the conversion data from a selected source is for measurement. In Ads Manager, open Tools, then Conversions, and select your data source to review its score and warnings. Those warnings can point to missing matching information, delayed server events, repeated event firing or too little activity for a configured goal.
OpenAI refreshes the assessment daily using seven complete calendar days, with time allowed for late events to arrive. After a fix, check the next real conversion first. The score can move gradually as older events leave the window. An unavailable score can reflect limited activity or pending processing. Use the warnings to guide a tracking audit and assess campaign performance separately with attributed results and business records. OpenAI’s event quality guidance.
How do you fix weak conversion signals?
Start with one completed purchase, registration or enquiry and follow it through your actual customer journey. The OpenAI Pixel sends events from the browser. The Conversions API (CAPI) sends them from your server. Check each path you use, especially if a payment provider, embedded form or booking tool handles the last step.
- Check when the event fires. Tie a purchase event to the confirmed order. Confirm that the event reaches the correct data source and corresponds to the campaign’s configured goal.
- Keep the click reference through checkout. Where permitted, carry the original OpenAI click reference (
oppref) through redirects and pass it with the server event when available. If you send an email or stable customer ID for matching, follow OpenAI’s formatting and hashing requirements and respect the visitor’s consent choices. - Keep browser and server copies aligned. Use the same Pixel ID, event name and event ID for both copies of the same action. Preserve the event ID and original action time on retries. Check queues and failed requests if a warning shows that server events arrive late.
Hypothetical example. A gift shop sees 12 confirmed orders in its store but only four attributed purchases in Ads Manager. First it checks whether the right data source received the purchase events. It then checks the selected goal, date range and attribution settings. The store counts confirmed orders from all sources, while Ads Manager credits eligible ad interactions. An attributed conversion can also take roughly 24–48 hours to appear. OpenAI’s conversion tracking troubleshooting guide.
Which conversion should a ChatGPT Ads campaign optimise for?
Choose the supported standard event that best represents the outcome you want and occurs often enough to evaluate. A retailer with reliable order tracking can optimise for purchases. A service business might select lead submission, then use its CRM to check how many of those enquiries became qualified leads.
OpenAI’s Conversion optimisation objective uses one standard event per campaign. Changing that event requires a new campaign. With click billing, delivery optimises for conversions following a click. With impression billing, eligible view-driven conversions can also inform delivery. Keep the event, billing method and reporting windows visible in your review so a change in reporting scope is easy to spot. OpenAI’s conversion-optimised campaign guide.
What is worth testing before peak holiday demand?
Test a change that answers a specific shopper question, such as whether the item will arrive by a confirmed date. Compare that message with your existing ad while keeping the offer and destination comparable. If you use product feed ads, check stock, prices and destination pages before adding seasonal products to the campaign.
SEM Lion would review spend, attributed conversions and cost per conversion alongside confirmed orders or qualified leads. Record the date of tracking, offer and creative changes. That record helps you decide whether an apparent improvement merits another test or coincides with a change in what the campaign can measure. OpenAI’s holiday campaign guidance.
How does SEM Lion use Lion AI to measure, learn and optimise ads?
Lion AI supports SEM Lion’s campaign review by bringing authorised advertising and analytics evidence into a shared workspace. The team measures performance using the relevant account, date range and source definitions, looks for a plausible explanation in the data, then uses that evidence to propose the next test. A marketer reviews the evidence and decides which changes to make.
For a ChatGPT Ads review, take EQS warnings and attributed conversions from OpenAI Ads Manager. Compare those figures with the connected analytics and business records your team is authorised to use. Lion AI’s client portal is in a controlled pilot, with connections and reporting subject to account permissions and feature availability. See how Lion AI works.
What can you ask an AI assistant to help you review?
Replace the bracketed inputs with aggregate or redacted data. Keep customer identifiers and credentials out of the prompt.
Trace an Event Quality Score warning
Use [exact EQS warning], [selected data source], [expected event], [redacted test results] and [time of the action]. Identify the first step we can verify, from the event trigger to receipt and ad attribution. Give the expected result and who should check it. Cite current OpenAI documentation and label anything the evidence leaves unresolved.
Decide whether to change the campaign
Review [aggregate Ads Manager export] and [order or qualified-lead totals] for [date range and time zone]. Our campaign uses [goal, billing method and attribution settings]. Separate delivery, measurement and offer hypotheses. Propose one test using [budget ceiling and review date], name the outcome that would justify a change, and identify where more data is needed.
Can advertisers in Qatar and Romania use Ads Manager?
As of 4 November 2026, OpenAI lists Qatar and Romania as eligible locations for the legal entity billed through self-service Ads Manager. Check the target market and the options shown in your account before committing spend. OpenAI’s Ads Manager availability list.
For first-time setup, read SEM Lion’s guide to running ChatGPT Ads. If you already have a campaign, ask SEM Lion about a ChatGPT Ads measurement review.