ChatGPT Ads Results: How to Review Campaign Performance

On this page
Give recent conversions time to appear
OpenAI says attributed conversions can take 24–48 hours to appear in Ads Manager. If you review a campaign every week, use a period far enough back for the figures to settle. Mark the dates you changed the budget, creative, destination or tracking. Compare periods with the same setup before you draw a conclusion. OpenAI’s reporting guide
Read from delivery to qualified outcomes
- Impressions and spend. Check whether the campaign ran enough to learn anything. If delivery is sparse, review campaign status, targeting and bid competitiveness.
- Clicks and site visits. Check whether people responded to the ad and reached the intended page. A click and an analytics session count different actions.
- Conversions and qualified outcomes. Compare the reported event with the action you actually wanted to buy. Use your CRM or order system to check which enquiries qualified and which purchases went through.
Ads Manager can combine conversion types in one total. To isolate an event, OpenAI documents this route: open the three-dot menu, select Edit columns, expand Conversions & events and choose the event. Export daily values if you need to compare the timing of changes. OpenAI’s guide to event columns and exports

Calculate the cost of an outcome that counts
Divide spend by the number of outcomes that meet your definition of success. If you need qualified leads, use the qualified-lead count from your CRM. Use the same cost basis across channels. Media spend alone cannot be compared fairly with a figure that includes management and creative.
If you had no qualified outcomes, report the spend and the zero. A cost per outcome is undefined in that case. For ecommerce, check whether the margin on attributed sales covers the cost of acquiring those sales. Revenue divided by ad spend measures return on ad spend, which is different from profit.
Pick the next change
- Little delivery: Check campaign restrictions before rewriting the ad.
- Clicks but few visits: Test the destination, redirects and analytics.
- Visits but few completed actions: Compare the ad promise with the page, then test the action yourself.
- Enquiries but few qualified leads: Compare the offer and form with the criteria sales uses.
- Qualified outcomes within budget: Check whether the result holds over a longer period before increasing spend.
Change one material variable at a time when you can. Record the metric you expect to move and the decision you will make when you next review it.
