I have talked to you about user operations before. What exactly is user operations?

Today I’m going to talk to you about the core of user operation: user stratification strategy in user operation.

In essence, it is a strategy similar to price discrimination, which adopts different strategies for different users, ultimately achieving the effect of enhancing user value and increasing user ARPU value.

This is why the prices of goods seen on different mobile phones and accounts are different. This is how big data killing old customers comes about! As a loyal old customer, there is no need to push more products with discounted prices to you. This may reduce the user value. Therefore, for old customers, the core of user operation is to find ways to make old users buy more and consume more, so as to contribute more value to the product.

Today we will briefly talk about user stratification in large-scale user operations. User segmentation is only the most basic and important step in refined operations. After segmentation, the strategy is continuously adjusted according to changes in user data, and the final formation of a set of data distribution logic is the core.

The price discrimination among different users, user levels and even points that we see are all different manifestations of user stratification.

1. The logic of user segmentation

User stratification is closely related to user profiling. Users are stratified and graded based on their basic attributes, user behavior attributes, user consumption characteristics, user life cycle and other characteristics.

For example, Weibo users can be roughly divided into content consumers and content producers. Among content producers, there are professional content producers in various fields. Among professional producers, there are divisions based on the number of fans. For example, content producers with less than 50,000 fans are called F1, and content producers with 50,000-100,000 fans are called F2. Customized operational measures are taken according to the user's level, including but not limited to resource support, exposure recommendation, event exposure, etc. A growth plan is formulated for them to ultimately achieve the goal of combining content production and demand, and promote the healthy development of the entire platform.

As for content consumers, each user will have his or her own exclusive label, and content will be recommended based on the user's content consumption habits. The platform will even actively push content to users, allowing them to continuously consume content on the platform.

On some content consumption platforms, users may be stratified and graded based on whether they have consumption behavior and how much the single consumption amount is. The ultimate goal is to encourage users without consumption behavior to consume and to encourage users who already have consumption behavior to consume more.

This is also the reason why Pinduoduo now uses methods such as hundreds of billions of subsidies to acquire users. As long as the user has his first consumption, we can obtain user data one more time, which makes it possible for continued consumption. The initial subsidy cost can be completely offset by the subsequent continuous consumption. If the initial user conversion cost is 100 yuan, then if the user can continue to consume on the platform in the future and become a loyal user of the platform, then this initial user conversion cost can be easily leveled out.

As mentioned earlier, users can be stratified based on a series of characteristics, but this has extremely strict data requirements. Perhaps only some large Internet companies with massive amounts of user data can present the entire user portrait.

For general Internet companies, user segmentation based on user behavior trajectories and consumption characteristics is the crudest user segmentation. The user's gender, age, and even region may not be easy to obtain, but as long as the user has left a footprint in the product, we can divide the user into different layers based on the user's behavior trajectory.

According to the user's login, opening, browsing, browsing time, consumption and other behaviors, we can simply divide users into 4 categories:

The first category: highly active and high-spending people. This type of users are loyal users of the product and are willing to spend time and money on the product. We need to serve this group of users better. Serving the 20% of highly active and high-spending people well may generate 80% of the product’s revenue, or even more!

The second category: high-activity, low-consumption group. Although this group of users has low consumption capacity, they are highly active in providing products. Although direct conversion is not possible, if we can use the diffusion ability of these users to increase the number of new users of the product, and use the fission ability of these users to bring exposure and new users to the product. Then the value of these users will also be reflected. Each type of user has its value, and we just need to utilize operational strategies and product mechanisms to maximize the user's value.

Category 3: Low-activity, high-consumption group. Although this type of users has strong purchasing power, they are more demanding on products and are not very loyal to products. Find what you want among multiple products.

Category 4: Low-activity and low-consumption population. This group of people has weak consumption capacity and is not active. Therefore, the most important thing for this group of people is to first improve one of the indicators through operational strategies, and then gradually convert it.

After user segmentation, the next step is to determine the operational strategy for each type of user, and to carry out targeted conversions through specific operational measures for each group of people.

2. Operational Strategy after User Segmentation

Once user stratification and grading are determined, products will be pushed to users based on the data presentation of different users and in combination with different resource positions.

There are two types of push logics. The first is the user operation logic that combines user stratification with resource positions and the operation actively configures resource positions for distribution. This is a common practice for most companies.

However, some problems may arise. That is, as the user scale increases, the number of divided user sets will also increase. At this time, if operations are required to perform configuration, the workload will be very large, and sometimes the time node may be missed.

At this time, the logic of data distribution is needed. Through changes in user behavior, users will be actively divided into different user groups, and then strategies can be written online to convert users through different operational strategies. This can not only save manpower but also improve traffic utilization efficiency.

The most typical example is the data distribution logic of Taobao that guesses what you like. Taobao will push the necessary products to each person in a customized manner based on changes in user behavior and search records. This is why people are willing to shop on Taobao. Because the logic here contains big data's behavioral positioning of each user, and it may know you better than your own mother!

3. How to continuously adjust strategies based on user behavior

Whether it is the operational strategy of actively configured user stratification or the operational strategy of data distribution, it ultimately needs to come back to the data. Find the starting point of the operation strategy from the data, and ultimately you need to go back to the data to verify the correctness of the operation strategy.

Infer the operational strategy through the changes in data in each period, and then make adjustments based on the data results.

When doing user operations, you must first stratify existing users according to one or several attributes, and then conduct the final conversion of users based on the performance of each type of users, in order to achieve the effect of increasing user ARPU value.

Wei Ya’s private traffic is operated by utilizing the user stratification model. Wei Ya will sort her fans according to their Taobao live broadcast levels and direct fans of different levels to different communities. For example, the "Weiya's Women" group requires fans to be at least diamond fans or true fans. These fans themselves have a certain stickiness to Wei Ya’s live broadcast, and will also be the most contributive group in future live broadcasts.

The most basic part of user operation is user segmentation. The most difficult part is to formulate corresponding user strategies for each user group through user segmentation strategies, and formulate strategies exclusive to each user segment.

Author: Wang Ting

Source: Operation Wang's Growth Diary (yunyingwang001)