Creating a Cohort Report
Suppose you want to see how long it took new visitors to complete a certain action for the first time over the past month. You can use the Cohort report to view that information.
By the Numbers
You can see in the report that of the 308 visitors who came to the site during the week of February 13th, 0.3% of them first performed the action 'View Item' within 7 days of their first visit, and no visitors during that week performed that action after the first week. However, for the week of February 20th, there were 4126 visits and 0.7% of visitors completed the "View Item" action. Within 14 days of their first visit, 0.1% of visitors completed the "View Item" action.
The various shadings of the cells (a heat map) is gives a quick visual cue to the general pattern of the cohort. The darkest color is assigned to the largest valued cell (1.0% in this example), and the rest are shaded on a linear scale proportional to that, with 0.0% completely white. This shading can help you to come to some quick conclusions without reading the numbers.
For instance, in this case we can see that the highest percentage of users tends to perform the 'View' action within a week of their first visit (as perhaps we expected). And that typically a small percentage of visitors are completing the "View Item" action, but that during the week of February 27th a larger majority of visitors completed the action.
This is an invitation to speculate about what occurred on the site that week: was there a new promotional message that went live during the week of February 27th that motivated visitors to complete the "View Item" action? And if this message was remove or altered the next week, it could explain why the March 6th group numbers returned to average for the site over that month. Cohorts are a great way to see patterns of behavior.
This page has no comments.