ChatGPT Ads Manager shows a panel called "Improve conversion performance" with three recommended steps for advertisers: improve conversion signals, optimize for customer actions, and consider automatic bidding. The panel appears in a screenshot shared in a LinkedIn post by Thomas Eccel, editor of The PPC Times, where it is marked as new. As of November 20, 2026, we have not found an OpenAI announcement of the panel, and OpenAI's Ads developer documentation does not mention it. The steps matter most to advertisers running ChatGPT Ads campaigns that still optimize for clicks.

What does the Improve conversion performance panel in ChatGPT Ads Manager show?
The panel sits beside the campaign table in Ads Manager. Each step has a short explanation and a button that leads to the related setting. In the screenshot, the first step has a green tick and the other two show an empty dashed circle, which suggests the panel tracks which steps an account has completed.
- Improve your conversion signals. The panel tells advertisers to "Send high-quality conversion data to give ChatGPT stronger signals". Button: Review event quality.
- Optimize for customer actions. The panel recommends conversion optimization focused on "purchases, sign-ups, or other valuable actions—not just clicks." Button: Review eligible campaigns.
- Consider automatic bidding. The panel suggests letting bids adjust within the budget for advertisers who do not know what to bid. Button: Review bidding settings.
The screenshot does not show what the Review event quality button opens. OpenAI's developer documentation, as read on November 20, 2026, does not describe an event quality score.
How do the three steps relate to OpenAI's Ads documentation?
OpenAI's Conversion Tracking page describes two ways to send conversions: the Measurement Pixel and the Conversions API. Advertisers can use both, and the page says to send "the same event name and event ID through both integrations so OpenAI can recognize the duplicate." For the Conversions API, it adds that supported customer information can be included to improve matching.
On optimization, the same page says a conversion-optimized campaign needs exactly one active standard conversion event setting attached at creation, and that the campaign objective cannot be changed after creation. OpenAI's Bidding & Budgets page adds that custom events cannot be optimization goals and that conversion campaigns bill for valid clicks.
For bidding, the Bidding & Budgets page describes Maximize Results, which lets OpenAI adjust bids toward clicks or conversions using the daily campaign budget.
Why it matters for advertisers
The panel puts conversion data quality first, before campaign settings. That order suggests OpenAI expects conversion-optimized campaigns to work better when the events behind them are complete and correctly deduplicated, though OpenAI has not said so publicly.
Because the campaign objective cannot be changed after creation, moving a click campaign to conversion optimization is likely to mean building a new campaign rather than switching a setting. Advertisers who follow the panel's second step should expect a learning period and a break in their reporting history.
The panel is an in-product recommendation, not a requirement. Accounts with too few conversions for one event may see unstable results from automatic bidding, so the advice could suit larger advertisers better than small tests.
What to do now
- Check whether the Improve conversion performance panel appears in your ChatGPT Ads Manager account. Availability may differ between accounts.
- If you send events through both the Pixel and the Conversions API, confirm they share the same event name and event ID, as OpenAI's Conversion Tracking page asks.
- Pick one standard conversion event that reflects real value, such as a purchase or lead, before creating a conversion-optimized campaign. Custom events cannot be used.
- Test a conversion-optimized campaign next to an existing click campaign rather than replacing it, and compare cost per conversion against your own backend data.
