App promotion: How to do user retention analysis? Make these two points clear!

There are two things you need to clarify before doing a retention analysis.

First, we need to identify the key events of the product . The so-called key events are when users feel the core value of the product. Only when users keep repeating the key events can we prove that the product is constantly creating value for users and forming a virtuous circle.

Secondly, what is the average period of users using the product ? This period directly determines the retention period selected when observing retention.

How to identify key events within a product?

When determining product key events, there are two aspects to think about.

The first is within the product, what we hope users will do when using it every day;

The second is one’s own business, such as which indicator can directly measure business development, or which one hopes for sustainable growth.

▌How to determine the life cycle of a product?

Through the above two ways of thinking, we can basically determine what the key events within the product are. At this point, if you clarify what the average user usage cycle within the product is, you can complete a basic retention analysis.

For example, in the past 60 days, 100 users may have triggered two key events in the product. These 100 people were selected to calculate the time interval between the first and second key events. After statistics, it was found that 3% of the people triggered it twice within one day, 20% of the people triggered it twice within 7 days, and on the 30th day, 81% of the users had triggered it more than twice. At this point, it can be basically determined that 30 days is the average usage cycle of users' actual feedback on the product, and it can be used as a cycle state of retention observed in retention analysis.

30 days may be the monthly retention. The monthly retention you see at this time may be a relatively meaningful analysis result.

▌How to calculate retention analysis?

Through the above steps, the key events to be analyzed can be clarified, and the retention period of the analysis can also be clarified. At this point, whether through tools or by writing data queries, you can get a table or a picture of the data retention.

Taking the above figure as an example, from September 12 to September 18, the 7-day retention of users in these seven days, we can calculate that there were 1,752 users on September 12, and on the third day, only 14 users came to visit again; at this time, the third-day retention corresponding to September 12 is 14/1,752, or 0.8%. By analogy, we can calculate the third-day retention for all dates, such as the 12th, 13th, 14th, 15th, and 16th, and then get the third-day retention of all users in the product. The actual retention rate might be 0.76%.

▌How to interpret the retention curve?

Through the above steps, you can draw the retention curve of the product for this analysis, which is a decreasing curve starting from the 1st day and then to the 7th day. Normally, product managers hope that the retention curve will remain as flat as possible after a period of time, rather than still being in a downward trend; because when the retention curve is relatively flat, it proves that the product has found a user group that matches the product in the market space, thereby proving the value of the product in the market.

On the other hand, when an enterprise carries out operational activities or makes product revisions and upgrades, it hopes that the entire retention curve will have an upward trend. An increase means that the overall retention rate is gradually increasing, which is equivalent to the subsequent user retention basis and quantity of the entire product, which is an improvement compared to before.

▌How to find the starting point to improve retention?

Once you get the retention curve, you also know how to interpret it. When we find that the retention curve is not flat or the overall retention rate is not particularly ideal, we need to analyze the reasons. Generally, we can do some retention analysis and comparison from five perspectives.

The first thing to do is to compare the retention of different user groups.

The second thing to do is to compare the retention of different features within the product.

The third is to compare the retention of users with different behaviors.

Fourth, we need to compare the retention situation in different dimensions.

Fifth, compare the retention of users who repeat key events at different frequencies.

Source: Analysys