How to use the pyramid model to implement user stratification operations (with case studies)

As the Internet becomes more and more permeable in our lives, the cost of acquiring new users for any product is increasing. Moreover, when the user base reaches a certain level, we have to pay attention to the operation of existing users. At this time, it is imperative to build a mature user operation system.

In all Internet products, each user plays a different role and performs different behaviors. Therefore, when we build a user operation system, we mainly stratify users based on their roles and behaviors.

This is also the premise of what we call refined operation in daily life. Only by doing a good job of user segmentation can we formulate corresponding operation strategies for users at different levels and achieve refinement.

Since we want to segment users based on the roles they play and the behaviors they perform, some questions arise: there are so many types of user roles, and user behaviors are even unpredictable, so how do we decide what form of user segmentation to use?

We can use the following common methods to make judgments:

1. Pyramid model layering
2. AARRR model stratification
3. User Value Model Layering

Today, we will mainly talk about the pyramid model layering.

Everyone should have seen the pyramid model, such as the one shown above: a pyramid model is built from bottom to top based on dimensions such as user activity, value contribution, and influence. Then, based on the user's performance in the above dimensions, different users are given different roles, powers, and promotion channels, forming a virtuous circle.

It should be noted that not all products are suitable for pyramid model stratification. Taking the alarm clock product we often use as an example, we cannot assume that a user is the most core user just because he sets many alarms every day. Products that use the pyramid model usually require the ability for users to influence each other.

The pyramid model seems to be the simplest of all hierarchical models. I think I can draw their model based on my understanding of a product. For example: Zhihu, those who answer more questions are core users, and those who answer fewer questions are non-core users.

Let’s not talk about whether this method of stratification based on experience and feeling is correct. Even if it is correct, how do you judge whether to answer more questions or fewer questions?

Therefore, the pyramid model itself is not as simple as imagined. It requires a full understanding of the business, users, and data to be successfully built.

After determining that a product is suitable for the pyramid model, first of all, we should sort out the core business logic of the product.

What is core business logic?

You can understand it as the functions that a product has in its earliest stage. If there are no major changes, it is the core business of the product. For this core business, the closed-loop relationship established by different roles within the product is the core business logic.

Take Maimai as an example:

After listing so many functions of Maimai, based on the product logic and business analysis, we know that the core functions of Maimai are finding jobs, building professional connections, and socializing in the workplace.

After determining the core business logic of the product, we need to think about:

• Will the outstanding performance, activity, and contribution of certain users create more value for the product?
• Is there a natural user growth progression without external interference?

We need to build a pyramid model based on the above thinking results. If the product will create more value due to the outstanding performance, activity, and contribution of some users, then we can stratify it according to user activity, scarcity, etc. If users will progress naturally, then build roles that allow for natural progression.

During the construction process, the naming of each level of users is relatively unimportant. What is important is how to define these user levels. For example: you can call the most core users in Zhihu "the leader" and the most popular users "the newcomer". These are all very simple. The key is how to define who "the leader" and "the newcomer" are, and find the nodes and basis for stratification.

• We can define the celebrities/ KOLs in Zhihu as "leading the trend". They are the key objects of Zhihu's maintenance, have a group of topic followers, are opinion leaders, and can produce high-quality and highly viewed content, thereby improving the overall traffic and interactive performance on Zhihu.
• We can define experts/high-contribution users in various topics/fields as "reaching the pinnacle". They can produce professional content, become a breath of fresh air in the era of self-media , maintain the influence of Zhihu's content, and answer the questions of users in need.
• We can define users who occasionally answer some interactive content (thanks, collections, praise), have a high usage of Zhihu, and love to praise and collect other answerers' content as "proficient". They are contributors to core traffic and disseminators of topic content.
• We can define ordinary browsing users as "newbies". They have not yet/do not like to answer questions, but just frequently query/browse some topics and do not leave too many interactive footprints. They are traffic contributors.

To find users at the above levels, we can use data, user tags, etc. I believe that most products have basic user data and behavioral data. For example: Chris is a celebrity in the field of Internet operations . He is very awesome and has joined Zhihu and started a topic. However, the number of his speeches and interactions are lower than the average. In this case, this celebrity may not be able to enter the "leading the way" level and needs more detailed stratified operations.

After completing user segmentation, we need to develop specific operational strategies for each level of users based on our product’s definition and expectations of those users.

Take the early Meilishuo as an example:

• Fashion experts are defined as "beauties who carry the brand image and external publicity." Mogujie 's expectation for fashion experts is to "establish a brand." Its core operating strategy is to "sign contracts and provide fashion experts with fixed income, image packaging and brand exposure."
• Super influencers are defined as those who "understand fashion, have the desire to express themselves, and produce high-quality content." Mogujie's expectation for super influencers is to "create content," and its core operating strategy is to "give super influencers a share of the profits and help them promote and increase their visibility."
• Active users are defined as those who "like fashion, may not be able to produce high-quality content, but are willing to spread and share." Mogujie's expectation for active users is to "spread content." Its core operating strategy is to "give regular feedback to active users, invite them to participate in offline activities , and send them small exquisite gifts."
• The demand masses are defined as those who “consume content and leave after finding what they want.” Mogujie’s expectation for the demand masses is to “consume content,” and its core operating strategy is “site-wide activities, distribution of various coupons, discounts, and vouchers to guide conversions .”

Once we have completed the above, we can do some specific operational work that can be implemented based on the core operating strategies for users at each level and in combination with the company's existing resources and means.