The relationship between ad density and session revenue is nonlinear, and most publishers don't map it correct

The relationship between ad density and session revenue is nonlinear, and most publishers don't map it correctly.

Common assumption: more ad units = more revenue. This holds up to a point. Beyond that point, ads push content below the fold, increase scroll distance, degrade user experience, and reduce pages-per-session. Your overall session RPM can actually drop even as per-page CPMs hold steady.

The metric to optimize is session revenue, not per-page revenue. A user who visits 4 pages with 2 ads each generates more ad revenue than a user who visits 1 page with 6 ads and bounces.

Run ad density experiments with holdout groups. Reduce ads on entry pages (high bounce probability) and increase them on engaged content (users who've already read 2+ articles). This changes the distribution of ad exposure toward users who were already going to stay, rather than users you might drive away.