Data operation: How to use data analysis to improve user retention?

Acquiring users is only the first step, but retaining users is the ultimate goal of all products.

On January 15, Duoshan, Matou MT and Chatbao were launched at the same time, sparking heated discussions about the " siege on WeChat ". However, a data report released by Getui Big Data showed that one and a half months later, the performance of these three applications was somewhat unsatisfactory.

As shown in the figure, driven by the press conference and heated discussions on the Internet, the three applications experienced explosive growth in users on the 15th and 16th. However, from the third day after the launch, the number of new users of the three applications began to plummet to almost the same level. Acquiring users is only the first step; retaining users is the ultimate goal of all products. Today we will talk about how to use data analysis to improve user retention rate.

1. Why do we need to retain our products?

1. Increase user growth rate and reduce customer acquisition costs

Retention has a long-term impact on active user growth. Here we specifically simulate a product growth situation. Suppose a product acquires 100 new users per week, and the next-week retention rate of new users is 60%.

In the figure on the left, the absolute value of the next week retention rate of users decreases by 3% every week, that is, the retention rate this week is 60%, the retention rate next week is 57%, and so on. In the picture on the right, the absolute value of retention rate decreased by 1%.

The weekly active users consist of new users this week and old users who have remained in the previous week. After 29 weeks, the previously retained old users in the left figure have basically decayed to 0, while in the right figure we can see that some of the earliest users are still able to stay after 29 weeks. Comparing the two situations, their new user acquisition volume is the same, and the retention rate is only 2% different in absolute value, but after 29 weeks, the ratio of user volume is 1:2. This chart vividly illustrates the importance of retention rate to the growth of active users.

In addition, in an interview with nine top growth hackers, Qu Hui, the author of "Silicon Valley Growth Hacker's Practical Notes", four out of nine chose retention as their favorite growth lever. Because retention can bring huge compound interest effects, and the cost ratio of maintaining old users and acquiring new users is about 1:5.

2. Retention is the most important criterion for judging product value

The retention rate actually reflects a conversion rate, that is, the process of converting initially unstable users into active users, stable users, and loyal users . As the statistics change, operators can see changes in users over different periods of time and thus determine the product's appeal to customers.

3. Understand the length of the user life cycle and optimize the product

Through retention analysis, you can check whether the new function has different retention effects on different groups after it is launched? Can you determine whether a new product function or an activity has increased the user retention rate? Taking into account many factors such as version updates and marketing promotions, we cut off functions with low frequency of use, achieved rapid iteration verification, and formulated corresponding strategies.

To sum up, if the product cannot retain users, our product is like a hollow basket. The more water we pour in, the more water will flow out, which means that our product cannot achieve sustainable growth. In addition, as traffic becomes increasingly expensive, retaining old users becomes increasingly important.

For users, the higher the retention rate, the better the product grasps the core needs of users , and the stronger the dependence of users on the product; for products, the higher the retention rate, the more active users the product has, the greater the proportion of users converted into loyal users, and the more conducive to improving the product's monetization capabilities.

2. What is User Retention?

In the business world, our ideal situation is to make the user's life cycle (those using the product) consistent with the product's life cycle , so that the product can develop in a healthy and sound manner, but this is not the case in reality.

In the Internet industry, because new customers are attracted by activities such as attracting new users or promotions, users start using the application within a certain period of time. After a period of time, users who still continue to use the application are considered to be retained users of the application. The ratio of these retained users to the new users at that time is the retention rate.

For example, a travel app gained 500 new users in July. Of these 500 users, 250 launched the app in August, 200 in September, and 150 in October. This means that the retention rate of this wave of new users in July is 50% after one month, 40% after two months, and 30% after three months. After talking about retention, let’s take a look at retention analysis.

Retention analysis is an analytical model used to analyze user engagement and activity levels , examining how many users who perform initial behavior will perform subsequent behavior.

3. Characteristics of User Retention

This is a common retention curve, which can be divided into three parts: the first part is the oscillation period, the second part is the selection period, and the third part is the stable period.

1. Oscillation period

We can see that the number of new users entering the website or downloading the APP has dropped dramatically in the past few days, from 100% to a few percent or even lower. This period is called the oscillation period. At this stage, we mainly focus on user activation . At this stage, we need to let users feel the core value of the product quickly and at low cost , and quickly reach the "aha moment".

2. Selection Period

After the oscillation period comes the selection period. Generally speaking, customers have a preliminary understanding of our products during this period and begin to explore our company's products to see if the product meets some of their core needs. At this stage, we focus on improving the retention of old users . We need to build the core functions of the product and cultivate users' usage habits for the product.

3. Stable period

After the selection period, there is a stabilization period, and the retention rate enters a relatively stable stage. At this stage we need to think about what the long-term value of the product is to users? How can we make users experience the value of the product repeatedly? In short, only by doing a good job of retention analysis and improvement at each stage can the retention curve be improved overall.

4. How to analyze retention data

1. Time grouping

The most commonly used method in retention analysis is to group users by time and then observe the changes in this group of users over time. As shown in the figure, of course this table can also produce the curve mentioned above. By counting the retention curve of users who are newly added, active, or have special behaviors and meet specific conditions on a certain day over a period of time in the future, we can summarize the time period from user addition to user loss and find the key link where the retention rate drops significantly.

At the same time, by comparing the subsequent retention changes of users in various channels, activities, and key behaviors, we found the factors that influence the improvement of user retention rate. For example, the retention rate of users who have received coupons is higher than that of users who have not received coupons.

Generally speaking, 30-50% of new users leave on the second day, and it is very common to only have 10% left after a week. In addition, many users leave without even using the core functions of the product. Based on this conclusion, in recent years, operational activities and product design directions have increasingly focused on providing care or onboarding tasks to new users in the first week, strengthening users' experience of the basic functions of the product. For example, first-time credit card swipe gifts, interest rate coupons for new customers of Internet financial products, etc.

In addition, we can also divide users from different channels into groups . After sorting the retention rates of all channels, we can easily find that some channels bring a large number of users, but the retention rate is very low. At this time, we can stop investing in these inefficient channels, which will save a lot of marketing costs.

2. Behavioral Grouping

Conduct one-on-one user behavior analysis for user groups with high churn/high retention, and count the behavioral characteristics of retained/churned users. In particular, for churned users , summarize the reasons for churn through behavioral analysis of churned users, thereby improving retention rate. For example, the behavioral characteristics of users who churn the next day after being added are that they open the homepage and browse for 30 seconds before exiting, and they do not enter the channel page and detail page. The characteristics of users who are retained the next day are that the first visit is more than 3 minutes and the browsing path is deep. Therefore, it can be judged that the reason for user churn the next day is due to lack of understanding of product capabilities, and user guidance needs to be strengthened.

In addition, user behavior is closely related to the functional modules of the product . We can also group the use of the functional modules of the product and optimize the functional modules.

According to the above ideas, retained user analysis can establish retention curves for different groups of users, observe the retention of users with different characteristics , and find out the influencing factors.

5. User Retention Analysis Process

Specific plans may be:

1. Define high-value retained users

2. Identify features or benefits that attract users to stay by sorting out data features

3. Analyze the access paths and habits of high-value retained users

4. Target high-value non-retained users

5. Through data features, identify the obstacles that high-value non-retained users encounter when exploring product features and benefits

6. Come up with targeted adjustment policies based on the characteristics of the obstacles

7. User portrait research to support analysis conclusions

6. Case Study: Improving New User Retention of Pan-Entertainment Apps

It mainly provides users with content consumption and has certain social attributes, so the daily active number and retained number of users are very important. Based on the current retention data of this product, we believe that there is still a lot of room for improvement in their new user retention. We want to find the focus of growth through data analysis, mainly by reducing the cost of getting started for new users and improving new user retention.

1. Understand the current status of new user retention

First of all, we need to understand the retention status of new users for analysis , because evaluating the status of a product is an unavoidable step before product optimization.

By comparing the retention curves of new visiting users and all visiting users, we will find that the retention rate of new users is significantly lower than that of all users. From this perspective, the retention rate of new users is a major growth point.

We just mentioned that for new users, the Onboarding activation process is very important . Whether users can quickly and cost-effectively perceive the value of a product when using it for the first time determines the activation rate of new users.

2. Determine activation goals

When talking about activation, we often talk about the "Aha Moment", which is the moment of surprise when users discover the core value of the product and generate the motivation for repeated use. The "Aha Moment" of different products is also different. For example, LinkedIn is to add 5 social connections in a week, and Facebook is to add 7 friends in 10 days.

We need to find the user's Aha Moment based on our own products and help users realize the product value as quickly as possible. First, we need to know which features may make new users feel the value of the product. We can use third-party data tools to analyze and compare the retention rates of users who have used different features to find the product features with the highest retention rate.

By comparing the retention rates of different functions, we can find the functions that have a significant increase in retention rate after new users use them. Then, based on the high or low cost of getting started with the functions for new users, we can further narrow the scope and determine the activation targets for new users.

3. Evaluate new user activation

After determining the activation goals, we can start to evaluate the activation status of new users and operate different users in a targeted manner. In this case, 30% of new visitors are not activated. For this part of users, what we need to do is to improve the activation rate; for the remaining activated new visitors, the optimization direction is somewhat similar to that of old visitors, that is, how to improve retention and let them continue to use the product.

The channel is the starting point of the user journey. New users coming from different channels will have certain differences , and their access intentions or goals may not be consistent. Therefore, we can look at the new user activation status of different channels. Here we can look at the differences between users from different channels from the perspective of new user activation ratio and retention rate.

It can be found that some channels have both low activation rate and retention rate, which may be because the channel population has a low match with the product demand. In this case, the channel needs to be adjusted more. Some channels may have a low activation rate but a slightly higher activation retention rate than the overall. In this case, the channel users may be more accurate, which requires further research on business data.

4. Activation time and retention impact

For inactive new users, when is the activation time more important? What impact do inactive users have on retention?

From the picture above, we can see that the activation time for new users should be as soon as possible. The green curve represents new users who were activated on the same day, and the red line below represents new users who were not activated on the same day. It can be found that the retention rate of the red line is very, very low. That is to say, if a new user does not complete the activation behavior on the first day of using the product, then the possibility of his loss will be very high, basically it is assumed to be lost.

So when we talk about activating new users, it basically means that the users can complete core behaviors and recognize the core value of the product on the day they enter the product.

5. Analysis of new user activation function

How to increase activation of new users within the product? Before asking this question, we can think about why new users are not activated and where they are lost? In this way, we can find the key steps of user loss and optimize them.

We can build a funnel based on the main path of new user activation . Through the conversion data of each step of the funnel, we can locate the main loss links of new user activation, and further analyze the reasons for loss by combining tools such as user segmentation and detailed investigation.

In the first major loss link, we use user segmentation to filter out new users who have not been activated in this link. By checking the behavioral trajectories of these new users in detail, we can find that more than half of the visitors are lost in the registration and login links. From the typical user behavior trajectory, we can find that many users are stuck at the login and registration step, and they have never touched the actual activation link.

So for the activation optimization of these users, we can advance the activation steps so that they can see the content as soon as they come in, without the need to register/log in.

In the second major loss link, we found that a small number of users have reached the list page, that is, they have reached the actual activation step, but they did not click on the specific content. Is it because the list content is not attractive enough to users? This list is composed of multiple contents such as user attention and product recommendations. It may be that there are fewer contents generated by user attention in the list. In this case, we need to better guide the attention behavior and increase the attractiveness of the content to users.

In response to this kind of loss, we made an adjustment to the product to advance the activation step, and found that the retention rate of new users has been significantly improved. But specifically, half of the visitors who had not been exposed to the actual activation steps have been successfully activated, and the retention rate of new users after the launch has also increased significantly.

VII. Summary

The purpose of user retention analysis is to summarize the characteristics of the user life cycle and find out the reasons for user retention/churn through refined analysis. In essence, it helps us understand the ability of our products to retain users and guides us to experiment, iterate and optimize our products. More importantly, retention analysis can help improve business models and decide on next steps.

Source: DataHunter (DataHunter01)