Ad refresh: measure the extra opportunity, not just the extra requests

A page generates more ad requests after refresh is enabled. That confirms the code is requesting more ads. It does not establish that the publisher created more valuable advertising opportunities.

Refresh replaces advertising content without necessarily reloading the page. Its value depends on the trigger, user activity, placement visibility, demand, and applicable partner requirements.

Google Ad Manager’s refresh-declaration guidance requires accurate declarations of refreshing inventory for the relevant setup. Treat that as product-specific guidance, not permission to refresh through a different product or partner.

Describe the trigger precisely

“Refresh enabled” leaves too much unspecified. Record whether refresh depends on an elapsed interval, a user action, a content event, or a combination of conditions.

Identify the minimum interval and what happens when the tab is backgrounded, the slot is outside the viewport, or the page is no longer active. Review the requirements of every demand partner involved before implementing the behavior.

The declaration and runtime code need to describe the same experience. A documented interval is not useful if another script refreshes the slot sooner.

Evaluate one session as a sequence

Imagine a reader spends several minutes on an article. The first ad is visible near the top, while later requests happen after the reader has scrolled far below it.

Counting those requests as equivalent opportunities would conceal the difference. Inspect whether the refreshed creative actually renders and whether it has a meaningful opportunity to be seen under the measurement method.

Keep initial delivery and refreshed delivery separate where reporting allows. That makes it possible to see whether later impressions have different viewability, demand, and revenue.

Follow revenue through the whole session

Suppose a simplified test increases impressions per session from two to three, while average revenue per thousand impressions falls from $3 to $2. Total ad revenue per thousand sessions is $6 in both cases, before other costs.

The extra impression count did not increase revenue in that example. It may also have affected page performance or the reader’s experience.

Compare revenue per comparable session or page view alongside impression-based metrics, traffic mix, and user experience. Use consistent definitions and retain the measurement coverage for viewability.

The viewability guide explains why a rate without coverage can mislead.

Give the test a stopping rule

Define the tested inventory, observation period, acceptable experience, and conditions for reversing the change. Monitor rendering failures, page stability, and any partner feedback.

Ensure a refresh does not replace an ad during an interaction or obscure content. The page-stability guide is relevant if slot dimensions or loading behavior change.

A successful refresh test should show additional value under the agreed rules. More requests are an implementation result; the business result needs evidence of usable delivery, revenue, and an acceptable experience.