This article will help you:
Identifying user behaviors that best predict retention can often be a long and frustrating process. Amplitude's Compass chart is designed to both speed that process up and make it more accurate.
One of the key steps in driving growth is discovering what your users' "a-ha" moments are. An "a-ha" moment happens when a new user makes the decision—consciously or unconsciously—to become an active user of your product. The most famous example probably comes from Facebook. They discovered that users who added at least seven friends in the first ten days were almost certain to stick around. Conversely, users who failed to do this almost always churned.
This revelation helped Facebook drive retention by encouraging the "right" user behavior: adding friends. Facebook could optimize their new user experience so that it emphasized adding friends over writing posts, uploading photos, or filling out their profile—all of which are actions that are central to the overall user experience, but don't contribute to long-term retention the way an early focus on adding friends does. And the fact that this metric—seven friends in the first ten days—comes at the top of the user retention funnel also gave Facebook the flexibility it needed to iterate quickly based on current data.
This is what the Compass chart can help you achieve. A Compass analysis scans through your user data and identifies these behaviors in moments, giving you the insights you need to efficiently improve your product and drive sustainable growth.
Before you begin
First and foremost, events will not appear in any Amplitude charts until instrumentation is complete, so make sure you've got that done. You'll definitely want to read our article on building charts in Amplitude, as this is where you'll learn the basics of Amplitude's user interface. You should also familiarize yourself with our helpful list of Amplitude definitions.
Also, this is the first in a series of articles on using the Compass chart. See the next one here.
NOTE: The Compass chart is only available to users on our Enterprise, Growth, and Scholarship plans.
Set up a Compass chart
If you're already familiar with Amplitude, the first thing you'll notice about the Compass chart is that it doesn't work the same way other Amplitude charts work. There's no Event Module, no Segmentation Module, and no Metrics Module:
When you first open a Compass chart, you'll see a heat map view that shows the likelihood of new users being retained into their second week, based on both the events they fire, and when they fire them. In this example, we see that users who start a session on their first day are less likely to be retained. This is probably because all new users will start a session on their first day—they have to start a session to become a new user in the first place, and since they haven't yet had a chance to interact with your product beyond that, it's not an especially predictive event.
But if they also edit their profile on that first day, they're much more likely to be retained into the second week: the correlation rates between firing those events on day 1 and sticking around for another week are 0.21 for starting a session, versus 0.47 for editing their profile. (We'll talk about correlation in the article on interpreting your Compass chart.)
To build a Compass chart, follow these steps:
- In the left-side module, select the user cohort you're interested in from the For users in the following cohort dropdown. This will be your base cohort. Amplitude will populate this list with the cohorts you've already created. If you have not previously created any cohorts, you will only be able to select New Users.
- From the ... how does performing dropdown, select an event you'd like to know more about.
- If desired, add properties to your event by clicking + where, selecting the property name, and specifying the property value you’re interested in. In the example above, the event property is '[Amplitude] Country,' and the property value is 'Germany.'
You can add multiple properties to your event.
- Next, take one of the following steps:
- Set the range of first use. In other words, tell Amplitude that your event must have been fired within a set number of days after the new user first used your product. The default is seven days.
- Tell Amplitude you'd like to group your results by a particular property, by clicking + group by. This will generate a report explaining how different values of that property affect the correlation between your base cohort and your target cohort. In the example above, that property is 'Device type,' and the resulting report will break the analysis down by different device types.
- From the ...predict they will be in the following cohort dropdown, select the target cohort you're interested in. Here too, Amplitude draws from the list of cohorts you've already created. You can also select from a handful of pre-set, out-of-the-box cohorts:
- [Amplitude] 2nd Week Retention
- [Amplitude] 3rd Week Retention
- [Amplitude] 4th Week Retention
- [Amplitude] 2nd Month Retention
Users will be included in these cohorts if they were new during the time frame of your analysis, and if they fired an active event in the week (or month) listed after they were new.
- Set the date range of your analysis with the date picker.
Make sure this time range is long enough to generate a good sample size, and far enough in the past to ensure that all users in your sample have had an opportunity to be retained. For example, if you are looking at '[Amplitude] 2nd Week Retention' as your target cohort, then you should make sure that the new users in your base cohort have been new for longer than two weeks.
Export your Compass report
Export the Compass report as a .CSV file by clicking Export.
Add your Compass report to a dashboard
To add your Compass report to a dashboard, follow these steps:
- If you haven't done so already, save your chart by clicking Save.
- Click + Add to. From the dropdown, select the dashboard you'd like to add the report to, or select Create a new dashboard.
Now that you've successfully created a Compass chart, find out what it all means—check out our Help Center article on interpreting your Compass chart.