Change Details

Building Funnel visualization and Funnel data exploration * high level timeseries view of key metrics (save / ready rate) * Highcharts timeseries component

Notes from James ============== Important metrics: * bounce rate: open up editor, cursor blinks, leave = (ready-abort + ready-end) / ready-* * attempt to save but do not succeed: not(saveSuccess) / saveAttempt-* * attempt to save and succeed: saveSuccess / saveAttempt-* * are ready but do not attempt to save: saveAttempt / ready-* * investigate difference in ptwiki vs. plwiki (or other pairs that are wildly differing on core metrics above) ** look at network connection quality, use location info (generic if specific not possible) * comparisons to wikitext editor ** rates + funnel * do not care about timing information this quarter * under starburst, list main paths through funnel graph with numbers / percentages ** init-ready-intent-attempt-success ** init-ready-abort * how does failure type correlate to eventual success * look into init-ready-end and double check those are aborts and not some other silly problem

Building Funnel visualization and Funnel data explorationNotes from James * high level timeseries view of key metrics (s============== Important metrics: * bounce rate: open up editor, cursor blinks, leave / = (ready rate)-abort + ready-end) / ready-* * Highcharts timeseries component* attempt to save but do not succeed: not(saveSuccess) / saveAttempt-* * attempt to save and succeed: saveSuccess / saveAttempt-* * are ready but do not attempt to save: saveAttempt / ready-* * investigate difference in ptwiki vs. plwiki (or other pairs that are wildly differing on core metrics above) ** look at network connection quality, use location info (generic if specific not possible) * comparisons to wikitext editor ** rates + funnel * do not care about timing information this quarter * under starburst, list main paths through funnel graph with numbers / percentages ** init-ready-intent-attempt-success ** init-ready-abort * how does failure type correlate to eventual success * look into init-ready-end and double check those are aborts and not some other silly problem