Greetings,
When we need a graph to analyze the behavior of a cohort from last week (I chose “week” as an example of specific time period) and compare it to the week before that, the only option is to create two cohorts using the same properties within two different “between” date ranges corresponding to those two weeks. If we want to analyze the same behaviors during next week, we have to come back to the cohorts and update the “between” dates. It can get very unpractical when several cohorts are being used: if we want to compare 8 different behaviors on a weekly basis, we end up with the need to update 16 different cohorts for a single graph.
It would be a great option to allow cohorts to be dynamic within specific date ranges. For example by being able to set them on last week or last month (or the weeks and months before...) and make them dynamic, allowing the cohorts to update themselves, moving from one period to the next, instead of being only dynamic for the past 30 days or limited to their “frozen” between dates.
I hope my idea is clear and understandable. If so, feel free to share your thoughts about it. 😄
Matrak
