User operations: 4 common user stratifications

Today we will move on to the second section of content, 4 common user segmentation methods and implementation. In this part, I will explain to you in detail how to use 4 common methods to segment users. In the previous content, we have introduced the essence of the user segmentation method. I think I must emphasize to everyone that there are two core factors for implementing segmentation well.

Two core elements of layered implementation

1. Users at different levels need to be able to be identified and distinguished through data fields or labels;

After we find a model or method, users at different levels must be able to be defined, distinguished and identified by data fields or some labels.

For example, we may simply define users into several categories, such as enthusiasts, ordinary users, and super experts. This sounds like a division, but if there are no detailed rules behind the division, and if we use some data, fields, or labels to identify and authenticate users at each level, then this stratification is essentially unrealistic, except for being a YY.

Therefore, the goal behind stratification must be to ultimately find the data and fields to distinguish the users.

2. The operating mechanism or strategy for each type of user is clear and stable.

In the example of Maoyan Movie mentioned earlier, if we are currently targeting a group of users who are movie enthusiasts, and we just imagine an operating mechanism in our minds, and we want to carry out activities centered on theater films for them, and if we only focus on this granularity, you will find that this operating mechanism is actually unclear.

Therefore, we need to implement it so that we can combine the top three or top five popular videos launched in that month and release them to users across the entire site once a month; we may even release some activities regularly based on geographic location, on a monthly basis.

At this granularity, you will find that such a mechanism definition will be clearer and more stable. We must have these two prerequisites: users can be identified, and the operating mechanisms and strategies for each type of user are clear and stable. Only then can our tiered operations be better. These are the two core prerequisites.

After talking about these two premises, let me introduce to you 4 common layering methods and their applications.

4 common stratification methods and their applications

Let's first draw out a basic reference coordinate, which can be understood as a variant of the four-quadrant model. The four-quadrant model is divided into four quadrants by a horizontal axis and a vertical axis, but this model adds a horizontal axis to the four-quadrant model. We will explain why this horizontal axis is added later.

So, what does this model mean? We actually want to use it to differentiate product types. What are the dimensions of differentiation? There are actually two major dimensions.

The first dimension is whether the degree of standardization of the main chain of the business is high or low. What does it mean?

For example, for some tool products, such as an alarm clock, the process and links of using an alarm clock are very standard for everyone. It is nothing more than setting a time first, and then the alarm will ring at that time. When the alarm rings, we cancel it and that's it. The process is very simple and highly standardized. This is a standardized solution.

Another part of the demand, such as reading, may vary depending on the region where the people are located, their age group, and their identity. For example, if the people are teachers, college students, or factory workers, the demands behind it will be very diverse.

Typically, this type of demand has a low degree of standardization and is mostly non-standard. For some non-standard demands, the business chain behind them may also be longer and more complex.

For example, the online education business has a very complex and diverse chain behind it. From the early enrollment to the user's entry into the class, to the maintenance during the class, to the user's feedback after completing the practical assignments, each link in the middle is a variable, and each link means huge uncertainty. In addition, people's learning needs and learning habits are relatively diverse, which makes this a relatively complex business.

So this is the first dimension. We need to judge whether the degree of standardization of the main chain of a business is high or low.

So what is the second dimension? That is, whether the possibility of users influencing each other in a product is high or low. Users of some products will establish relationships with the product, but users of some other products may not. Even for products of the same type, the likelihood of users influencing each other may vary.

For example, in a product like Toutiao, the possibility of users influencing each other is relatively low, but in Zhihu the intensity of user relationships is significantly higher, and users will have more opportunities to influence each other. For products like P2P and financial management, users' demands will be highly personalized and non-standard, but it is almost impossible for users to influence such products, and we do not need users to see or touch them.

Once we have such a model to determine where a product should be located, we can then respond in four ways. The four most common layering methods are suitable for what kind of products. Let’s first take a look at what the four methods are.

1. User Personalization and Demand Segmentation

The first method is to segment and stratify users based on their personalized characteristics or personalized needs. This stratification method is more suitable for products with low standardization in the main business chain, diverse personalized needs of users, and more complex businesses. This is the first stratification method, and its specifics will be discussed later.

2. User identity segmentation and stratification

The second method is to segment and stratify users based on their identities. This stratification method can be applied to products where users are more likely to influence each other.

In a product, if users can influence each other, then you can give users a stronger identity, such as adding a V, giving them an honor or a medal. This will make it more meaningful. In other words, its creation can lead some people to feel admiration and envy for this thing. These emotions can then generate a stronger driving force in the product and guide users to perform certain behaviors.

There are two remaining stratification methods, which can be used by almost all products without making any distinction based on product characteristics. These two stratification methods are:

3. User value segmentation and stratification

We stratify users by judging whether their value is low or high, and use this as a reference.

4. AARRR model stratification

This is a model we often see in the context of growth. Through this model, we can also complete a rough stratification of users and make some targeted operational strategies. You can define it as a growth model or as a user stratification model.

The above are the four types of stratification methods that I will talk about with you one by one. Before I officially start explaining how to use each type of stratification method and how to finally implement it in the data to identify and segment users, let’s take a look at it again.

I just talked about a coordinate, which may seem a bit abstract to some people, so let's take a look at a few examples. Let's take a look at several products, including Moji Weather, Amap, Evernote, Duoduo, China Merchants Bank, and TikTok. If we follow the reference coordinates mentioned above, where will these products be placed? Let's use this example to familiarize ourselves with the reference coordinates.

1. Moji Weather

First of all, for Moji Weather, we still look at it from these two dimensions. In terms of the degree of standardization of the main business chain, it is a weather forecast product, and users check it mainly to obtain weather-related information. The degree of standardization is high because the weather conditions seen by everyone should be the same, so the main chain at least has a high degree of standardization.

It simply means that depending on the weather conditions in different regions, we may be able to push some information to everyone, and there may be some possibilities for refined operations in this area, but if we only look at the main chain, the degree of standardization must be high.

The second dimension is whether the likelihood of users influencing each other in the product is high or low? There is no doubt that for a product like Moji Weather, the possibility of users influencing each other in the product is obviously relatively low. Users do not need to establish relationships in it, and do not need to interact a lot in it. So basically, if we only consider the main chain, Ink Weather can be placed in such a position.

It meets both the main chain’s high standard and the low possibility of users influencing each other.

2. Amap

We also return to these two standards. In the case of map navigation, the main chain of Amap's business, the degree of standardization seems to be relatively high, because the navigation solution should be consistent for all users, and there will not be a large number of personalized needs. Secondly, the possibility of mutual influence is also low, so it should be placed in the same position as Moji Weather.

3. Evernote

Evernote is a note-taking tool product. The standardization level of its main business chain may be relatively low, but the personalized needs may be relatively high.

Because a large number of users may have different usage and needs for the same note-taking tool. Some people may use notes to make their daily to-do lists and manage their own affairs, while others may use Evernote as a knowledge management tool to store a large amount of information. Anyway, the usage is different, and the support provided by Evernote for each type of usage will also be different. Some people may also use Evernote to share some data documents with colleagues, so there will be some differentiated things.

Therefore, the degree of standardization of the main chain of the business may be lower, but there is no possibility of users influencing each other. After all, Evernote is still a tool, and the interactions between users are invisible. Therefore, Evernote should be placed at this position of the coordinate. The main chain has a low degree of standardization, and users do not influence each other.

4. Get

It’s different for getting. First of all, the degree of standardization of the main business chain is relatively low for things like knowledge, acquiring knowledge, or acquiring some audio content to learn. There is no doubt that each type of user is very different, so the degree of standardization of its main business chain will be relatively low.

At the same time, with a product like this, users are actually visible to each other in the product, especially after using some community-type products, users can see each other, so the possibility of users influencing each other in the product is relatively high. Therefore, the degree of standardization is low, but the possibility of influence is high.

5. China Merchants Bank

China Merchants Bank is a bank financial investment product. When it comes to financial investment, the user's degree of personalization is definitely relatively high, but it is almost impossible for users to influence each other within the product, so China Merchants Bank and Evernote are quite similar and will be placed in the same position.

6. TikTok

Douyin is a community-based short video product. First of all, users’ needs are definitely more personalized and differentiated. At the same time, users will definitely influence each other in the product. Therefore, if a product like TikTok is to be judged on this coordinate, it should meet both criteria at the same time.

If we were to make a judgment on these six products, based on their positions, we can basically determine what kind of user segmentation method they are suitable for. For example, for Evernote, it is more suitable for the differentiated features of users' personalized needs. And for Get and Douyin, it is definitely a stratification method that is suitable for both the personalized needs of users and the stratification of users according to their identity.

Products like Moji Weather are not suitable for the two stratification methods on the right. They are suitable for the two categories below, so this is basically the idea.

Author: Xiongxiong Operation Notes

Source: Xiongxiong Operation Notes