Hi, I have created a retention dashboard and something does not seem right. For the beginning event I have used 'create task' event. Also I have used retaining event as 'create task' which is the same. So I foresaw that the retention for the first period should be all 100% because both event are the same. However I am seeing metrics like 80% on the service. So can you explain what is going on?
Hi Dan H., Please stand by as I search the Mixpanel knowledge base for an AI-powered answer. I normally take about 15 seconds or so. In case AI cannot answer your question, feel free to reach out to Mixpanel support at mixpanel.com/get-support. If you have follow-up questions for this thread, please tag me again with your question: @kapa.ai <follow-up question> 👋
Feel free to ask a follow-up question once I've answered by mentioning me with @kapa.ai
In Mixpanel's retention analysis, the retention counts are based on unique users, not event totals. This means that even if the "beginning event" and the "retaining event" are the same, the retention rate might not necessarily be 100%. Here's why: The retention analysis in Mixpanel is based on user cohorts. A user cohort is a group of users who did the "beginning event" within a specific time window. For example, if the "beginning event" is 'create task', then a user cohort might be all users who did 'create task' within a specific day or week. The retention rate for a user cohort is then calculated based on how many of those users did the "retaining event" within subsequent time windows. So, even if the "beginning event" and the "retaining event" are the same ('create task' in your case), the retention rate might be less than 100% if some users did not do 'create task' again within the subsequent time windows.For example, let's say you have a user cohort of 10 users who did 'create task' on a specific day. If only 8 of those users did 'create task' again within the next day, then the retention rate for that user cohort for the next day would be 80%.You can find more details about how Mixpanel calculates retention in the Mixpanel documentation.
kapa.ai is still learning and improving, please let me know how I did by reacting below.
⤵️ Hint: Mention @kapa.ai in the thread for follow-ups.
No it does not make sense because, unique user count should be same for the first period (at least) if we are using the same events
Let me check the knowledge base for you again. One moment! 📚
I understand your confusion. However, it's important to note that Mixpanel's retention analysis is not just about whether a user performed the same event again, but also about when they performed it.When you set up a retention analysis in Mixpanel, you specify a "birth event" and a "retention event", as well as a "birth interval" and a "retention interval". The "birth event" is the event that defines the user cohorts, and the "birth interval" is the time window for defining these cohorts. The "retention event" is the event that we're interested in seeing if users perform again, and the "retention interval" is the time window in which we're looking for this event.So, even if the "birth event" and the "retention event" are the same, the retention rate might not be 100% if some users did not perform the event again within the "retention interval". For example, if the "birth interval" is a day and the "retention interval" is also a day, then a user would need to perform the event again within the next day to be counted as retained. If they performed the event again, but not until two days later, they would not be counted as retained.This is why you might be seeing a retention rate of less than 100% even though the "birth event" and the "retention event" are the same. It's not just about whether users performed the event again, but also about when they performed it.You can find more details about how Mixpanel calculates retention in the Mixpanel documentation.
kapa.ai is still learning and improving, please let me know how I did by reacting below.
⤵️ Hint: Mention @kapa.ai in the thread for follow-ups.
Let me check the knowledge base for you again. One moment! 📚
I understand your question better now. If the "birth event" and the "retention event" are the same, it would indeed seem logical that the retention for the first period should be 100%. However, the calculation of retention in Mixpanel is a bit more complex and depends on the specific time intervals you've set for the analysis.In Mixpanel, if you don't specify an interval or a unit, the interval defaults to 1 day. This means that each user gets 24 hours in each interval to do the specified event. So, even if the "birth event" and the "retention event" are the same, the retention rate for the first period might be less than 100% if some users did not do the event again within the first 24 hours.For example, let's say a user does the "birth event" ('create task' in your case) at 8 PM on a certain day. If they don't do 'create task' again before 8 PM the next day, they would not be counted as retained for the first period, even though they might do 'create task' again later. This could result in a retention rate of less than 100% for the first period.You can find more details about how Mixpanel calculates retention in the Mixpanel documentation.
kapa.ai is still learning and improving, please let me know how I did by reacting below.
⤵️ Hint: Mention @kapa.ai in the thread for follow-ups.
