By mastering these 4 steps, 90% of novices can operate "super users"!

In the second half of the Internet, what you need is not only user growth, but also continuous growth in value!

This article shares four points:

1. Defining superusers
2. Get superuser
3. Developing super users
4. Preventing super user loss

Step 1: Define a superuser

Super users are not the same as VIP users. VIP users come from traditional industries and often measure user value with "money". But in the second half of the Internet, in addition to contributing "money", users' greater role is to help products expand their user base and stimulate user demand.

We use four dimensions to measure super users: recognition (starting behavior), engagement (percentage of time spent, percentage of spending), stability (user profile, continued activity), and diffusion (active sharing). These four dimensions form a cycle in themselves. Only with recognition will there be investment, continuous and active dissemination, and greater recognition. What is considered here is not the absolute value, but the proportion of investment.

From the perspective of the user life cycle, we can quickly distinguish "churned users, silent users, and novice users" to locate high-value users in the growth stage, and then find highly active and high-contributing users, i.e. super users, among the high-value users:

1. Through the linkage of Push products, find lost users and users who uninstalled the App;
2. Find silent users through user insights;
3. Target novice users from the perspective of user investment;
4. Ultimately identify high-potential users and continue to cultivate and develop super users.

The above functions can all be realized through the AI ​​intelligent new customer acquisition solution provided by U-App AI version, which can find potential super users in the industry; find the distribution of potential super users in various media through cooperative media, and formulate targeted delivery strategies; and then use AppTrack mobile advertising monitoring to aggregate population data, media data, and delivery data in the AI ​​model to further optimize the selection model and solve the problem of delivery strategies being heavily dependent on experience and cost. This series of processes is a joint optimization between us and the media.

Increasing the overall number of super users can be achieved through three means: acquisition, development, and retention. Convert as many users as possible who are about to leave into super users, guide all new users to become super users, and continuously enhance the value of super users.

Step 2: Get a superuser

There are two dilemmas in obtaining a superuser:

• First, the cost of acquiring customers is getting higher and higher. For some industries, the cost of acquiring customers for Apps is as high as 300 yuan per person, which is a very scary number.
• Second, delivery optimization relies heavily on experience, and delivery operations on various media channels are complex.

How to crack it?

Before launching campaigns, know who the potential super users are and launch campaigns in a targeted manner. This requires the use of AI capabilities to select super users and distribute them across various channels. For example, there are 10,000 potential super users, 8,000 of whom are in media A and 2,000 in media B. We can achieve more results through targeted delivery, and then monitor and optimize in combination with post-link behaviors.

To sum up, there are four steps:

1. Find users: find potential super users for advertising;
2. Determine strategy: Place ads according to the media channels of the target group;
3. Look at conversions: Locate the super users who ultimately convert through ad monitoring;
4. Optimal delivery: Intelligently optimize delivery plans based on user clicks.

If we can find potential super users in the industry through AI models and global user portraits; find the distribution of potential super users in various media through cooperative media, and formulate targeted delivery strategies; and then use mobile advertising monitoring to aggregate population data, media data, and delivery data in the AI ​​model to further optimize the selection model. This link can solve the problem of heavy reliance on experience and cost in delivery strategies. This series of processes can be linked with the media and optimized together.

If you want to develop super users, you need to find potential users. But the huge differences between industries make finding potential users elusive.

Let me share a real case. A client was operating two apps at the same time. One was a reading app, which he started with; later it expanded to a mother and baby app. In the figure below, the horizontal axis represents time and the vertical axis represents the degree of investment. right

For reading apps, the competition is very fierce, with some gaining while others losing; for maternal and child apps, users spend about the same amount of time on each app, and each app has its own focus and strengths. Because the identity of the audience is different, mothers will use several apps at the same time. Therefore, the two industries have completely different portrayals of super users.

This link can find potential super users in different industries through "time series model", "segment modeling" and "training structure based on recurrent neural network".

Step 3: Develop super users

When you find a super user, you need to know what he likes, cater to his likes, and convert him into a super user.

To know what someone likes is to understand the person.

• One is interest preference . What kind of person is he? What characteristics does he have? These are mostly user labels and portraits.
• The second is behavioral preference . The difficulty in capturing user behavior preferences is that for an App, there are two types of behaviors: front-end behavior and back-end behavior, and most of them belong to two sets of data systems. On the front end, user browsing behavior is monitored through embedded points, while the back-end data is stored in its own database. It is difficult to view user behavior data and transaction characteristics from a bird's eye view.

For example, connect the front-end behavior and the back-end behavior, and then measure the changes in users through event analysis, funnel analysis, and retention analysis. After identifying the user, you need to cater to his or her preferences. Combine interest preferences with behavioral preferences and develop targeted products to develop operational plans.

Of course, you also need to use industry benchmarks to understand the level of your user group in the industry. If the super user data improves but the overall user situation does not break through in the industry, it is likely that there is something wrong with the scope of the definition of super users.

Step 4: Prevent super user churn

After we acquire super users and develop more super users, we need to consider how to prevent them from leaving.

According to data from a US market research organization, the average number of apps downloaded per person per month in 2017 was 5.1, and the average number of apps installed per person per year in 2017 was 30. This shows that the probability of apps being uninstalled is very high.

To prevent loss, you need to have three capabilities:

1. User groups with high churn risk. The system automatically selects users with high churn risk. You can save them as groups for outreach and get to know the health of the App and user stickiness.
2. Choose your own risk threshold. Choose the risk threshold according to the App’s development stage and industry attributes;
3. Complete risk distribution display. The loss warning model is displayed in a distributed manner to ensure good early warning and prevention work.

The most direct way to win back users is Push. There are many strategies for Push delivery. For example, the two Push delivery methods above have many problems in terms of push timing, text description, and audience selection. If Push notifications disturb users too much, users will choose to turn off the App notifications or even uninstall it.

So, recovery is not about disturbing others, what you should do is to know yourself and your opponent.

You also need to predict who the potential users are and what their behavioral habits are; in Push, whether the user opens the App at the right time, you need to receive timely notification, and Push must be timely. After the push is sent, we need to understand the user's post-link behavior to find out which users are converted into super users, and finally form a complete closed loop.

Based on DI's global data intelligence + AI capabilities, operators are empowered to explore and realize the five data capabilities of monitoring, analysis, attracting new users, management, and reaching out. They measure who will become super users from four dimensions and define super users from the perspective of the user life cycle. Understand user behavior through user insights, and know and cater to user preferences through global user tags; win back silent users through Push, and take operational precautions through churn warnings.

In addition, combined with user management, from a life cycle perspective, all lost and silent users can be turned into potential super users, and novice users can be converted into super users.

Source: Umeng Global Data