3 Operational Guidelines for APP Product Loss User Research: Why Do Users Leave?

Before actually contacting users for research, the following issues need to be addressed:

●Prevent user churn or? Regain lost users? ——Different purposes require different questions to be answered. Only by recognizing the purpose can the solution be implemented;

●Who are the lost users? ——Accurately define lost users to avoid obstacles caused by vague concepts;

●Design a systematic diagnostic framework - Establishing a diagnostic framework and logic tree guidance can greatly improve the efficiency of problem solving.

1. "Preventing user churn" and "recovering lost users" are two completely different goals

The research on the reasons for user churn must ultimately be implemented on the solution. Different purposes require different questions to be answered in the research:

●Prevent user churn: Why do users churn? (Reasons for user churn) → Will users continue to churn? (Are the dissatisfaction points of existing users consistent with the reasons for churn?) → How to prevent user churn? (What improvement measures should be taken);

●Recover lost users: Why do users churn? (Reasons for user churn) → What is the possibility of lost users returning? →Under what circumstances would they be willing to come back? (Return touchpoint) → How to recover lost users? (What improvement measures should be taken).

2. Who are the lost users? -Use the return visit rate inflection point method to find lost users

If your goal is to prevent customer churn, when is the latest time to intervene? When users just churn out; if your goal is to win back lost users, when is the earliest time to take action? When the user just lost. As long as you find out "when users just churned", you can find out which users have really churned and the reference time for intervention.

We made a mistake here: because the team was operating a product that was paid on time and the KPI was final revenue, we initially simply and crudely defined the user churn standard as when the user did not renew their subscription. In fact, for this type of product, users have already left when they have not consumed all the products but no longer use or visit them. Such standards greatly delay the time we can intervene in lost users. Therefore, lost users should be users who have not performed key operations within a period of time.

●Choose key actions: Generally, the key actions are login or access. However, in our research on paid email products, we found that users will stop logging in soon after they stop sending emails. Sending emails can be used as another key action for reference.

●Choose the length of time: The length of "a period of time" varies for different products. For example, for community products it may be one month, but for e-commerce products it may be 3 months. So how do we reasonably define and judge the length of time that users churn?

To define the length of time that users churn, it is necessary to introduce the concept of "returning users". The value of user return rate can indirectly verify the rationality of the churn definition. Returning users refer to users who do not perform any key operations within the churn time limit but revisit after churn. We can calculate the user return rate, that is: user return rate = number of returning users/number of lost users*100%. Under normal circumstances, if the definition of churn is reasonable, the return visit rate of mature products should be below 5%~10%, and that of emerging products may be even lower.

The length of the churn time interval is inversely proportional to the revisit rate. We can determine the reasonable churn time interval by observing the correlation curve between the churn period and the revisit rate and selecting the inflection point within a reasonable range. As shown in the figure:

3. Design a systematic diagnostic framework

Designing a systematic diagnostic framework requires careful and detailed identification of research questions and formulation of hypotheses.

● Define the survey subjects - which type of users’ reasons for loss should be investigated: Do you want to investigate the users whose loss increases sharply within a certain period of time? Or is it to investigate one category of users whose churn is significantly higher than that of another category? Or conduct surveys throughout the user life cycle.

● List the hypotheses on the reasons for loss: The analysis method can refer to the MECE analysis method of McKinsey pyramid model, etc. The data attributes and log analysis of lost users can be of great help here.

What we do is to investigate the reasons for user churn throughout the entire life cycle, drawing on the analysis method of a simple analysis of the reasons for user churn in online games: comparing the current status of users using the product with the summarized churn types, and analyzing the user's demand content and churn influencing factors at different stages.

1) Define the user life cycle

The user's life cycle is divided into four stages based on the user's usage time, number of logins, number of payments, etc.: contact and adaptation period, exploration and growth period, maturity and pursuit period, and fatigue and decline period.

●Users in the contact adaptation period are in the initial stage of using the product, from the first minute of contact to the first exit. During this period, users need to adapt to and slowly accept the product's visual style, operating procedures, content, etc. The contact period is completed when the user logs in for the second time.

●The exploration and growth phase will vary in length depending on the user. During this process, users begin to become familiar with the functions and design of each aspect of the product, and accumulate more and more content, habits or social relationships. If there is a complete new user growth system, it will be easier to divide the time.

●The mature pursuit period is the period when users enter a relatively stable state. Users will consume and accumulate more content, form more regular usage habits, adopt more in-depth and convenient operating methods, establish more contact relationships, etc., becoming the core user group of the product.

●Users in the fatigue decline stage gradually lose the desire to use the product, key usage behaviors gradually decrease, and they basically do not perform more difficult operations or tasks in the product. If a product cannot even allow users to spend their time at a very low cost, it is very easy to cause churn.

2) Summarize the types of churn

The hypotheses on the causes of attrition can be roughly classified into four types: rigid attrition, frustrated attrition, natural attrition, and market environment attrition.

● Rigidity loss: refers to factors that are difficult to avoid or require major design changes to adjust. The reasons for rigid loss of different users are different, which are related to product positioning, functions, and visual design. For example, strict spam control can easily lead to the loss of users who receive promotional emails; another example is when users who like fresh and fresh styles see cool and tough designs. Rigid churn is most likely to occur in the early stages of user use, and the churn rate is relatively high.

● Frustrated churn: This is an important reason for churn during the contact and exploration periods. The factors causing churn are complex but easy to discover and remedy. When users feel discomfort beyond their tolerance during use, they will become frustrated and leave. For example, encountering content that is extremely difficult to understand, functions with inaccurate guidance, or the failure to provide convenience after investing a large amount of usage costs, etc., will cause users to become bored and disgusted.

●Natural churn: refers to the situation where users find that there is nothing to watch or lose interest in product features after using the product for a period of time, have no goals to pursue, and enter a period of fatigue, resulting in churn. Natural loss is the best reflection of product design level and the greatest test of product vitality. For example, the core content is single, lacks living space for different users, and cannot capture user interest; another example is that the product's peripheral systems are not rich enough (points, memberships, etc.), and cannot play a supporting and coordinating role.

●Market environment: Churn is often difficult to control. It is not just a problem with the product itself, such as the impact of OTT products such as WeChat on text messages and phone calls.

3) Phased loss comparison

Compare the characteristics of users at different life cycle stages with the types of churn, and propose hypotheses about the causes of churn at different periods. Here is a simple example:

Now that the preliminary work is completed, we can start designing the investigation plan. Do you feel that the vague questions have become clearer now? To develop the most effective way to collect the required information, you need to determine the following: data sources, survey methods, research tools, sampling plans, and of course, budget applications, etc. As for whether to reach users through phone calls, text messages, emails, field visits or search engines, use 5Why tools, competitor comparisons or questionnaire card tests, etc., it all depends on personal opinion. Each has its own strengths, so just choose according to the actual situation. The rest depends on the specific product, so I won’t go into details.