The Retention report in Mixpanel is designed to assess user engagement over a specified period of time. Retention of your users is critical to achieving product-market fit and sustainable long-term growth, and this report will help you understand how long users continue to come back and find value from your product.
Imagine your product is a B2B messaging application. You might use Retention to answer these sample questions:
- On average, how many users are still active after two weeks from signing up?
- What percent of all users are still sending messages after seven days?
- How has my 7 day messaging retention changed over time?
- What percent of users sent messages in 2, 3, or 4 distinct hours of the day?
Building your First Report
Building a Retention report follows the same high level steps as building any other report. You can read about report building basics here.
Building a report in Retention takes just a few clicks, and results arrive in seconds. Let's build a simple report together. Continuing the B2B messaging example, imagine you wanted to answer the following question:
Which app version promotes the best retention after signup, for users on iOS?
Step 1: Choose Events
In this case, since we want to know retention after signup, choose "Sign Up" in the first event slot. Then, we want to know if users are coming back and using the product, so choose "Any Event" in the second event slot. At this point, your query should look like this:
Step 2: Choose Filters
Filters exclude unwanted data. In this case, we only care about events performed on the iOS platform. Therefore, add a "Platform" filter, where Platform equals "iOS". At this point, your query should look like this:
Step 3: Choose Breakdowns
Breakdowns segment data into groups. In this case, we want to break our report down based on the app version they are using to Sign Up. Therefore, add an "App Version" breakdown. At this point, your query should look like this:
Congratulations, you've constructed your first Retention query! Now, it's time to examine the results.
By default, Retention displays a single, multi-part chart type to visualize retention data. It displays retention data as both a line chart, and a table. Data is identical between the two, but while the line chart provides a visual representation of users becoming inactive over time, the table provides a heat map to show which groups have the best retention. In this case, we can select Retention Table to see which App Version yielded the best results. Your report should look like this:
The report also offers a Retention Trends visualization to observe any changes in a particular retention bucket over time.
Now that you've constructed the query, and chosen the ideal chart type, you can easily answer the original question. Finally, save the report for later via the controls at the top right of the report.
Select Trends from the report drop-down list to see how the percentage of retained users has changed over time for any of the retention time unit buckets. You can change between the different time unit buckets (i.e. <1 Day, Day 1, Day 2, ...) to see if your retention metrics are improving or declining along your retention curve.
User Cohort Buckets
Retention counts users; not event totals. In other words, each of the user cohort buckets will include every unique user that did the "A event" criteria in that time window, starting at 0:00 of the first day of the bucket and ending midnight of the last day. A customer can only be counted once per bucket, but can be included in more than one bucket.
The first column (Date) indicates the day/week/month when the user performed the "A event". The Size column indicates the number of users that performed the "A event" within the time period.
For example, if you are bucketing based on your "Item Purchased" event and creating weekly buckets, a customer who purchased at least one item each week will be in every bucket, not just the bucket of their first purchase.
Each box within a row is assigned a shade of purple. The shading gets darker the higher the retention percentage. It's important to note that the scale is relative to each cohort row.
Boxes with an asterisk (*) indicate that the data is still in flux and not set yet because the time is still ongoing. Hover over a box to see when the last qualifying date for that bucket will occur.
While this article covers the basics of the Retention report, you can learn more about its capabilities in the following additional articles: