Are you perhaps seeing the effect of differential distribution of browsers across nations rather than differential distribution of user behavior across browsers? (Quick test: filter down to US traffic only, then look at the numbers again.)<p>If the US uses Firefox more than everybody else (plausible in my experience, but check your own numbers) then I'd expect Firefox would have markedly higher CPMs associated with it. Ditto Safari -- sure, somebody using Safari is probably "a Mac users ready to spend money, click on ads, etc.", but more importantly somebody using Safari is not Indian, Chinese, etc. (Picking big nations with low CPMs, not picking on anybody.)