Setting a price rule once and leaving it is a common mistake. Demand shifts by season, by advertiser category, and by traffic source — which means the 'right' floor price for a given ad unit changes more often than most publishers actively test for.
Why experiments matter here
A floor set too high suppresses fill and hands the impression to a leak. A floor set too low leaves revenue on the table on impressions that would have cleared at a higher price anyway. The only way to find the actual optimal point is to test it directly against your real demand, not assume it from a rule of thumb.
How to structure a price rule experiment
- Pick one variable to test at a time — usually the floor price for a specific ad unit or inventory segment.
- Split similar inventory into a test and control group, rather than changing the price for all of it at once.
- Run the test long enough to account for day-of-week and demand fluctuations — typically two to four weeks minimum.
- Compare realized revenue and fill rate together, not just one metric in isolation.
Reading the results honestly
A price increase that raises CPM but tanks fill rate can still be a net loss in total revenue. The only number that matters at the end is total revenue generated by the segment being tested, not any single metric along the way.
How AdPulse helps
We build and monitor price rule experiments as part of ongoing account management, so pricing keeps adjusting to real demand instead of sitting static for months at a time.
Want hands-on help applying this to your account?
Book a free consultation