Comments on: Cohort Analysis in RJMetrics http://themetricsystem.rjmetrics.com/2009/09/09/cohort-analysis-in-rjmetrics/ A blog about data by RJMetrics Wed, 03 Mar 2010 23:03:16 +0000 http://wordpress.com/ hourly 1 By: jakestein http://themetricsystem.rjmetrics.com/2009/09/09/cohort-analysis-in-rjmetrics/#comment-1412 jakestein Sun, 27 Sep 2009 16:38:47 +0000 http://themetricsystem.rjmetrics.com/?p=1021#comment-1412 Josh, the Y axis in the charts above represents revenue in a particular quarter as a percentage of revenue in the first quarter for the same cohort. For almost every cohort analysis we've done or seen, the subsequent periods show a decline from the first period (less than 100% when taken as a percent of the first period). To give you a numerical example, let's say that Vandelay Industries gained 100 new customers in Q1 2009. Those customers spent an average of $5 per person in Q1, yielding $500 of revenue for the 1st quarter of the Q1 2009 cohort. In Q2 of 2009, only a subset of the 100 customers from the Q1 2009 cohort purchase again, maybe 50 people. Unless the 50 customers who purchase again spend at least twice as much per person ($10), the value for quarter two for the Q1 2009 cohort will be less than 100%. Josh, the Y axis in the charts above represents revenue in a particular quarter as a percentage of revenue in the first quarter for the same cohort. For almost every cohort analysis we’ve done or seen, the subsequent periods show a decline from the first period (less than 100% when taken as a percent of the first period).

To give you a numerical example, let’s say that Vandelay Industries gained 100 new customers in Q1 2009. Those customers spent an average of $5 per person in Q1, yielding $500 of revenue for the 1st quarter of the Q1 2009 cohort. In Q2 of 2009, only a subset of the 100 customers from the Q1 2009 cohort purchase again, maybe 50 people. Unless the 50 customers who purchase again spend at least twice as much per person ($10), the value for quarter two for the Q1 2009 cohort will be less than 100%.

]]> By: Josh http://themetricsystem.rjmetrics.com/2009/09/09/cohort-analysis-in-rjmetrics/#comment-1411 Josh Sun, 27 Sep 2009 08:19:04 +0000 http://themetricsystem.rjmetrics.com/?p=1021#comment-1411 I'm not entirely sure what the Y axis here represents. What is the quarter 2 percentage in terms of? It seems like perhaps it would be in terms of cohort revenue from the previous period, but obviously that can't be the case since it would generally be greater than 100% for all future quartes. Can you give me some insight? I’m not entirely sure what the Y axis here represents. What is the quarter 2 percentage in terms of? It seems like perhaps it would be in terms of cohort revenue from the previous period, but obviously that can’t be the case since it would generally be greater than 100% for all future quartes. Can you give me some insight?

]]> By: LRH http://themetricsystem.rjmetrics.com/2009/09/09/cohort-analysis-in-rjmetrics/#comment-1386 LRH Wed, 09 Sep 2009 22:24:33 +0000 http://themetricsystem.rjmetrics.com/?p=1021#comment-1386 This is a great explanation of the value and complexity of cohort analysis. CYou clearly demonstrate the value of the RJM process and technology This is a great explanation of the value and complexity of cohort analysis. CYou clearly demonstrate the value of the RJM process and technology

]]>