Despite the world-changing events over the last few months, retaining your users remains as important as ever. The summer break is the perfect time to brush up on your retention graph skills\u2026 If you have a mobile application, you know the efforts that go toward engaging and retaining your users are considerable. Therefore, your investments must be profitable over the long term<\/strong>. If your visitors love your application, they\u2019ll use it many times over a certain period (and ideally this period goes for as long as possible\u2026). To measure their engagement over time, look no further than the retention graph.<\/p>\n A mobile app\u2019s success can be blindingly impressive if we only look at certain metrics (like the app download rate, for example). But downloading an app and leaving it on your phone doesn\u2019t mean you\u2019re using it regularly.<\/p>\n To study a mobile app\u2019s life cycle, it\u2019s interesting to study the retention rate over an observational period that\u2019s relatively long (several days or weeks) and then segment your data into several different visitor cohorts<\/a>.<\/p>\n According to \u201ctraditional\u201d calculation methods, the retention rate corresponds to the proportion of users who came back to an application on a given day (without taking into account any potential interactions between the user and this application before<\/span> <\/strong>this specific day). There are other calculation methods (like \u201cfull retention\u201d, \u201crolling retention\u201d, \u201creturn retention\u201d, among others) which we won\u2019t go into here. The \u201ctraditional\u201d method is the most widespread and gives a good idea of the application\u2019s overall retentive abilities.<\/p>\n With the retention graph (also called a retention matrix), you can compare different populations at a glance.<\/a> Good to know\u2026<\/strong><\/p>\n To read a retention graph, it\u2019s important to understand that the time-related element (the period analysed) is represented horizontally as well as vertically. Here\u2019s an overview of the main elements to understand:<\/p>\n Let\u2019s take an example\u2026<\/strong><\/p>\n In the following retention graph, we can analyse a period of 6 weeks, during which different user cohorts downloaded and launched an application for the first time.<\/p>\n Even if the absolute number of Cohort A users (those who first launched the app in Week 1) is higher, fewer of them came back later, compared to users from Cohort B. If we look at the week P+2, we see that 53% of Cohort B users came back to use the application, whereas only 38% of Cohort A users came back.<\/p>\n Why does the graph have this upside-down stairs shape?<\/strong><\/p>\n Each week (or each day, depending on the selected granularity), a new user cohort is analysed. As time progresses (the more recent the cohort is), the remaining analysis period automatically decreases, giving this graph the inverted stairway shape.<\/p>\n Digital analytics tools, such as AT Internet\u2019s app analytics<\/em> dashboards<\/a>, allow you to create your own graphs for retention analysis. To learn more, watch our video about retention matrices:<\/p>\n
Article originally published on 13th<\/sup> April, 2016.<\/em><\/p>\nWhy measure your apps\u2019 engagement capabilities?<\/strong><\/h2>\n
How is retention rate calculated?<\/strong><\/h2>\n
Enter the Matrix\u2026 <\/strong><\/h2>\n
![\"\"](\"http:\/\/s.atchik-services.com\/generateur\/img\/bird_gray_16.png\"){kind=icon}
![\"\"](\"http:\/\/s.atchik-services.com\/generateur\/img\/bird_blue_16.png\"){kind=icon}
<\/span>
Choose the appropriate granularity (day or week) and study recurring usage. You\u2019ll be able to evaluate the impact of a specific factor in particular (such as a campaign, an event, exposure to content, etc.) on application use.<\/p>\n![\"Main](\"https:\/\/blog.atinternet.com\/wp-content\/uploads\/2016\/04\/retention-graph-en-1.png\")
<\/a><\/p>\n![\"Retention](\"https:\/\/blog.atinternet.com\/wp-content\/uploads\/2016\/04\/retention-graph-en-2.png\")
<\/a><\/p>\n