i love data, but I hate how everyone in this sub obsesses over CPM like it's the holy grail. So my team and i pulled a year of aggregated data from over 1,200 publisher sites-sessions, RPS, fill rate, viewability, ad density-and the results pissed me off in the best way.
Turns out, the single strongest predictor of RPS? Not CPM. Not fill rate. Not some shiny viewability metric. It's impressions per pageview. Correlation was 0.59-nearly double anything else we measured. That means the number of ads that actually load on a page is way more important than what those ads pay per thousand. Volume drives the bus, folks.
and here's another one that wrecked my assumptions: session duration vs RPS had a correlation of -0.03. Basically nothing. But pageviews per session? 0.27-nearly ten times stronger. so if you're sitting there optimising for time-on-site, you're probably wasting your energy. push people to click through more pages instead.
the floor price trap is real and it's quantifiable. Publishers who right-sized their floors filled double the inventory and earned 19% more RPS, even though their CPM was 2.5x lower. Aggressive floors almost never pencil out-stop being greedy.
Someone in the thread asked about user retention and bounce rate when ad density gets pushed higher. great question-we didn't dig into that specifically, but it's the obvious next step. if you're ramming more ads down people's throats and they bounce faster, does that long-term trade-off kill your pages per session? I'd love to see that data.
full write-up is free if you want the methodology, but honestly, stop chasing CPM and start counting impressions per pageview. That's the real tea.
