How to find the reasons for decreased conversion by splitting “user active status”?

If you can clearly split and analyze the user's active status, then about 70% of data analysis problems will be solved. What often "stuck" the analysis is this very basic but easily overlooked content. Therefore, when you encounter any questions about data fluctuations, first segment the user activity status and clarify the activity status of the "users causing problems", which will definitely be of great help to you.

Number of new users + number of old users = number of active users

Let’s talk about the word “active” first. When we talk about active users, many people think that so-called active users are users who are very active in the product. To quantify it, for example, users who use the product at least two or three days a week are considered active users.

Sorry, that’s not the case, so first of all, we need to clarify a concept.

The so-called active, or active user, is generally defined in the industry as a user who has opened the product within a selected time period and is considered active, or an active user. So, activity defines a state, not a degree .

Active users are divided into two categories: new users and old users.

I won’t explain the new users, but the so-called old users, that is, users who are not visiting the product for the first time, are all old users. So the relationship between these three concepts is that in the same time period, the number of new users + the number of old users = the number of active users.

For example, the new active data you see every day, such as: an average of 4,000 new users per day and 10,000 daily active users, means that there are an average of 6,000 old users visiting every day.

Lost users + silent users = inactive users

Well, since we are segmenting the user's active status, if there is an active stage, there must be an inactive stage. If you pay attention to inactive users, you may be slightly surprised that the number of inactive users is extremely large.

Inactive users are also divided into two parts, namely, lost users and silent users.

The absolute majority of them are churned users. The so-called churned users are those who have used our products but have not launched the products for a period of time. Moreover, this period of time is so long that we believe that the users have disowned or forgotten the products. Then we define such users as churned users. According to the business characteristics of different products, they are generally divided into 30 days, 60 days, or more than 90 days.

The other part is silent users. Similarly, silent users have used our products before, and have not launched the products for a while. However, this time period is a time interval with maximum and minimum values. The maximum value cannot exceed the value that defines lost users, and the minimum value is generally one-third of the number of days defining lost users.

Okay, here is a key point. Many people will ask me how do you judge or define lost users. My answer is that this threshold is defined based on our understanding of our own business and users and is gradually calibrated through data. There is no official formula.

As your product grows, the number of inactive users may be much larger than you think. It is very important to recall inactive users, and it is very effective if done properly.

Because of this, some users will become silent users or lost users and then be successfully recalled, becoming a very unique group among old users, called returning users, or returning old users.

Why do we need to do such segmentation? Because the usage scenarios and experiences faced by a returning user are very similar to those of a new user. We also need to activate returning users and keep them active. However, they are not new users in essence, so they cannot enjoy preferential treatment such as financial products and novice labels. Therefore, it is necessary to segment this type of users and provide exclusive operations and services.

First, when a user enters our product as a new user, there will be two directions:

1. If successfully activated and recognizing the value of the product, the new user will continue to visit and become an active old user;

2. If a user has not visited the product for a period of time after being added, he or she will become a silent user. When the period of time during which the user has not visited the product reaches the standard for lost users, the user is in a lost state.

At the same time, if a user who is in a silent or lost state visits our product again due to our recall strategy or because he saw our advertisement or thought of us when he had a need, such a user is in a reflux state. After the user returns, if he continues to visit, he will become an active old user.

Finally, if an active user, whether he is a new user, a returning user or an old user, he may become a silent user at any time. This is why data is needed to monitor the status of users in the product in real time so that strategies can be adjusted in a timely manner.

First, from the traffic dimension, we usually only focus on the number of new and active users; if we look at the change graph of user activity status, you will find:

☞The newly added status is the starting point for any user.

☞Silent state is the inevitable path for a user to go from active to lost.

☞For any recall strategy, the user will definitely experience the state of reflux.

Therefore, new addition, silence, and return are the three key nodes of the entire user status. Everyone pays enough attention to new users, but silent and returning users are often easily overlooked. Therefore, when we do traffic analysis, we must be able to accurately measure new user acquisition, activation and recall.

The analysis of new users is to attract new users and promote activation

The analysis of returning users is to promote activation after recall

The purpose of analyzing silent users is to prevent and recall them in time

Secondly, from the conversion dimension, the focus of our analysis should be on the people who truly influence conversion. Many companies will encounter the dilemma of "why the conversion rate has decreased, and I can't find the reason". In fact, we need to know that the users are the ones who are most likely to be changed by data fluctuations. If the data indicators you check every day do not break down the user's active status, for example, once the quality of new users decreases, the conversion rates of all your key indicators will decrease. Therefore, when checking the key conversion rate, it is necessary to segment it and segment the conversion rates of different user statuses, such as: the conversion rate triggered for the first time; the conversion rate triggered repeatedly by old users.

Third, from the retention dimension, although we always mention the retention rate, in fact, 90% of people have a very shallow analysis depth on retention. If we want to expand on the content related to retention, it involves user life cycle calculation, cohort analysis, etc. I suggest that if you want to understand the value of segmented user status, then at least, in addition to retaining new users, you should also measure the retention of returning users and old users so that you can clearly evaluate the operational results.

Activity is a state, not a degree. We need to segment user activity status , meet the needs of users in different statuses and encourage them to complete conversions. We need to accurately measure user activity status and formulate strategies for attracting new users, promoting activation, and recalling users. We need to evaluate the user's value level, identify the groups that truly influence conversions, and measure the retention indicators of active users at each stage. Only with such refined operations and improved satisfaction of users in different activity statuses can we ultimately achieve business growth.

Source: Zhuge io Data Coach