When ad revenue falls, compare matching report totals before changing ad settings. Fill rate describes one part of delivery; viewability describes a different measurement. Either rate can change without proving the cause of the revenue movement.
Start with the same reporting scope
Choose the same dates, time zone, currency and inventory filters. Write down revenue, ad requests, filled requests or impressions, measurable impressions and viewable impressions. Keep the report’s original metric names beside the numbers.
Google’s metric reference contains different fill-related definitions for different reports. Do not assume every rate uses the same numerator. Our fill rate calculator uses the filled-request count and total-request count you supply; that must match the comparison you intend.
Keep two calculations separate
For a fictional request-based example, 18,000 filled requests out of 20,000 total requests produce a 90% ratio. Separately, 6,000 viewable impressions out of 10,000 measurable impressions produce 60% viewability.
These denominators are deliberately different. Do not divide 6,000 by 20,000 and label that the same viewability measure. Use the viewability calculator with matching measurable and viewable figures.
Why a higher fill ratio can coexist with lower revenue
Consider a second fictional period. Period A has 100,000 opportunities, an assumed 80% filled proportion and $4 per thousand resulting impressions: $320. Period B has 50,000 opportunities, 90% and $3: $135. This simplified model assumes one resulting impression per filled opportunity and matching revenue scope.
The filled proportion improved while volume and rate fell. The arithmetic illustrates why one percentage is not an explanation for total revenue. It does not diagnose a real site or prove why any auction price changed.
Investigate one observation at a time
- Break down the change by device or placement before editing the whole site.
- For a delivery issue, inspect the actual page and platform reports. The Google Publisher Console is an official GPT diagnostic tool.
- For visibility questions, review real measurement and page behavior. A locally calculated percentage does not measure ad exposure.
- Record a specific proposed change and the result you expect to observe. Compare like-for-like data afterward.
The Active View overview explains that served impressions may not actually be seen. Avoid treating all impressions as measurable or assuming that placement alone determines the result.
Build a practical follow-up list
Use the monetization health report to organize matching-domain scan evidence and your own metrics. Missing evidence should remain a question, not become a score. Confirm conclusions in the relevant reporting system.
Should I add more ad slots immediately?
A lower total is not enough evidence for that decision. First identify which counts or rates changed and consider the actual reader experience before testing a layout change.
Official references
References checked: 26 October 2026. Examples identified as fictional are teaching inputs, not performance benchmarks.