I agree that the common presentation of cohort data is clunky. That's why I built this tool with d3js to visualize cohorts as stacked bars: http://bslatkin.git

I agree that the common presentation of cohort data is clunky. That's why I built this tool with d3js to visualize cohorts as stacked bars: http://bslatkin.github.com/cohorts/ http://bslatkin.github.com/cohorts/ "Impact maps" are too blurry in my experience; they assume your traffic is normal and predictable. Cycle plots are fun to look at, but in practice the lines are too volatile to make a pretty chart.