Segment your users and how to conquer them one by one?

Today, we will talk about practical techniques for user segmentation, which is an essential skill for finding the most critical core users among many users, and for fine-tuning design and operations .

There are many angles of classification. Let’s first review what kind of classification is good from a logical perspective? My understanding is that after dividing the entire set into subsets, the differences between individuals in different subsets should be as large as possible, and the differences between individuals within each subset should be as small as possible. Then, our principle for classifying users is:

For users in different segments, the differences in “demand scenarios” should be as large as possible.

1. If the product has multiple users, first classify them according to different roles

Multi-sided products, corresponding to single-point (such as widgets ) and unilateral (certain like-minded social applications) products, must have at least two clearly different user groups, and usually such products have platform attributes.

For example, for Taobao, the common classification is buyers, sellers, third-party service providers, and platform parties; for Didi Chuxing, it may be divided into drivers, passengers, and platform operators; for Zhihu, it can be divided into questioners, answerers, spectators, and platform parties... This is a classification based on different roles. Generally, we will use different roles as the first dimension to classify target users. The demand scenarios of different user groups are obviously very different.

2. Newcomers, intermediate users, and experts

This is a classification based on the user's "familiarity with the product field" and is also a very commonly used user classification method. For one-sided user roles, if you can't find a better way to divide them, I suggest using this method as a backup. After all, the demand scenarios of novices and experts are very different. The former hope for "simple, easy to use and easy to get started", while the latter expect "stable, reliable and high performance".

For example, on QQ Music, for users who listen to music, if they are new, they may need the hottest charts and automatic song selection through scenes; for expert users, they may search for a certain album or a certain singer; for example, when driving, a newcomer needs an automatic transmission that is easy to drive, while an expert may need a manual small steel cannon that can go on the track or a hardcore off-road vehicle that can go into the mountains.

A similar perspective is new users and old users.

3. Based on demographic information

Age, gender, occupation, location, consumption level... This method should be used with caution to avoid situations where demographic information has little to do with the product (for example, differentiating taxi users by occupation is illogical). The differences in demand scenarios for the several categories of users divided in this way are often not very obvious.

Let me give you a few examples that are more appropriate. Bank credit cards of various levels are divided according to assets, income, and consumption level; Lego toys are divided according to age—the normal intelligence level of children of different age groups; the style strategies of some clothing brands are divided according to region—first-tier cities and third- and fourth-tier cities; milk powder, different age groups; students in K12 education, different age groups; toilets, gender (this example is ridiculous, haha)... and so on.

4. Based on the business scenario of the product

It is difficult to summarize this into a general rule, so let's give you some inspiration through a few examples, such as:

• For enterprise services, it can be divided into seed round, angel round, A, B, C... PreIPO, listed, etc. according to the stage of the user (company);
• For training, you can divide them into industries such as the Internet, IT, fast-moving consumer goods, and manufacturing, and different positions such as products, operations, technology, and management;
• For air travel, it can be divided into ordinary, silver, gold and platinum cards according to the user's accumulated mileage;
• Media people can distinguish readers based on their three perspectives: believers in traditional Chinese medicine and opponents of it, nationalists and those who believe in world peace;
• Those who build communities are often divided into PGC writers who create original content, activists who process content (like, forward, etc.), and browsing users who simply consume content;
• Travelers can be divided into singles, couples , parents and children, team building, etc.

There are also more general ones, which are divided into rookies who need to be guided, loyal fans who need to be maintained, old loves who need to be recalled, and annoying people who need to be driven away;
…

Make up a doggerel (what the hell, it doesn't rhyme at all)——

Multilateral first divides the sides,

Newcomers and experts,

Demographics,

Business scenario.

If you don't divide it well, it will be a mess later. If you divide it well, it will be easier to divide and conquer.

Mobile application product promotion service: APP promotion service Qinggua Media advertising

This article was compiled and published by @苏杰 ( Qinggua Media ). Please indicate the author information and source when reprinting!