HI I don't get why this squares getting smaller. On 1st week we have filled data to the bottom and every next week is 1 square less. Could you explain this ? Btw Great video it is not boring, it seems dynamic and I understand much more than from other videos :D
+Michal Tom That's because we have more month over month data for the cohort that joined our service the earliest i.e. in January in this case. The February cohort was not active in January, March cohort was not active in February and so on. Hence you see that each subsequent month will have 1 less square (i.e. data) than the previous month.
+Hipokryta Tom I'm assuming you're talking about how in September 2010, for example, there's data until 12 months in that particular row, but then every subsequent month, it decreases by one month. If that's your question, I guess that's because users who signed up in a later month (say, march 2010) haven't been around as long as those who joined earlier (september 2010). The "User signed in by months" field basically shows how much percentage of users who signed up in that particular month sign in after a certain amount of months. So the data in the 6th column for september 2010 (3.7%) means that 3.7% of people who signed up in september 2010 still signed in 6 months later. Let's suppose that this analysis was done in september 2011. When we're looking at data for september 2010, we have data for the past 12 months, because september 2010 was 12 months ago, and you can see how many people from that "cohort" have continued to sign in each month. But if you were to look at October 2010, that was only 11 months ago, so you can see how many people who signed up then for the past 11 months. For each subsequent month, the amount of months the users have been "retained" (been around) will decrease by 1. Hope this pointlessly long comment helped :)