Interpret your stickiness analysis

Stickiness will help you dig into the details of engagement across your user base, specifically regarding how many users have formed product usage habits.

This article will explain the Metrics Module of the Stickiness chart, and will help you interpret your stickiness analysis.

Before you begin

If you haven't done so already, be sure to read our article on building a Stickiness chart in Amplitude.

Interpret your Stickiness chart

In Amplitude, Stickiness can be measured in one of two ways: cumulatively or non-cumulatively. You can change this setting at the top of the Metrics Module at any time during your analysis.

Non-cumulative stickiness

The non-cumulative Stickiness chart shows you the percentage of users who fired the event at least once on the exact number of days listed on the X-axis. For example, users in the 2 days bucket have fired the event on exactly two days over the course of a week (or month) in the time frame of your analysis, while those in the 3 days bucket have done it on exactly three days in a week.

In this example, over 70% of users who added friends during the last 12 weeks, did so on only one day during a given week in the analysis's time frame. Roughly 24% did so on exactly two days, while no users did so on all seven days of any week in your time frame.

Instinctively, it might seem like a user can only appear in one bucket of a non-cumulative stickiness analysis. This is not the case. A single user can have different stickiness values for each week (or month) in your analysis. For example, they might fire the event on one day in week one, and then three times in week two. This user would then be included in both the 1 day and 3 day buckets of your non-cumulative Stickiness chart.

Cumulative stickiness

The cumulative Stickiness chart shows you the percentage of users who fired the event one or more times on at least the number of days listed on the X-axis. For example, users in the 2 days bucket have fired the event on two or more days over the course of a week (or month) in the time frame of your analysis, while those in the 3 days bucket have done so on three or more days in a week.

Notice the percentage of users who have added friends on one or more days of a given week is 100%. This will always be the case in a cumulative stickiness analysis. That's because any analysis includes only users who have actually fired the event—and since, by definition, they all will have fired it on at least one day during our selected time frame, they'll all be included in the 1 day bucket.

You can also click on a specific data point to inspect the users included in that point. See our Help Center article on the Microscope for more information.

Breakdown data table

The table shows a detailed breakdown of the data by each user cohort and more granular daily buckets. In this example, there were 290,149 users in the week beginning October 17th. The 2 days cumulative stickiness is 76.8%, meaning that 222,834 of the 290,149 users fired an event on two days or more that week. Days with incomplete data will have an asterisk.

Track changes in stickiness over time

You can also discover how the stickiness of your most engaged users fluctuates over time, by selecting Change Over Time from the ..shown as drop-down menu:

In this example, we can see how two-day, three-day, five-day, and seven-day stickiness fluctuated with each week's new cohort of users. As we can see, they're mostly steady, with a noticeable twitch between October 12 and October 25. The product manager might want to investigate why this happened.

Create a cohort from your Stickiness chart

Users on Scholarship, Growth, and Enterprise plans can create a cohort of the users who make up a particular data point, via the Microscope.