How to create user portraits and find accurate users?

Two nights ago, a young girl who had just started working in operations came to me for help. She said she was in the jewelry industry and it was too difficult to increase the number of followers on her official account. Her leader asked her to build a community , but she didn’t know how to find her target users . I asked her what the user profile of your product was, and she said she was a 35-year-old woman. When asked again, she said with a confused look on her face that there was no more.

It's no wonder that she is in a difficult situation, because she has no idea who her customers are, where they are, or what they want. So how can she find them? It's like falling in love. Only by knowing the other person's thoughts and needs can you please her.

With the increasing variety of products today, the description of user portraits is crucial for corporate products. In the early stages of each project, the user labels of the products need to be very clear. Without portraits, there are no users. Only with precise positioning can you quickly find your target users and push your products to them.

1. What is a User Portrait?

User portrait is the labeling of user information. Analyzing various indicators of different groups, gaining insight into the core characteristics of user groups, restoring the true paths of users, and effectively outlining the user's behavioral portrait can often enable companies to make clearer product positioning choices and thus meet the consumption needs of target users.

2. Why describe user portraits?

User portraits can be understood as real virtual representations. They are not specific people or things, but rather are element representations of a unified type that are summarized, refined, and processed based on the product's target audience and its differentiation. It represents the audience and target group of the product.

User portraits can label and visualize user needs. Once the user portrait is formed, there will be huge improvements in data analysis and statistics, computer processing, product analysis , advertising , message push, etc. The user accuracy will be greatly improved, thereby improving work efficiency.

3. When should user portraits be created?

You can do it at every stage, but the requirements at different stages are different.

It can be roughly divided into two stages:

1. In the early stages of the product, when we didn’t have many users and data.

2. In the middle and late stages of the product, we have accumulated a large number of users and data.

In the early stages of the product, we did not actually have a very clear user profile. We relied more on our market research or the experience of relevant leaders and ourselves to judge user groups and user characteristics. At this time, our user portraits are not supported by a large amount of data, and we focus more on scenario analysis rather than user portraits. Analyze the scenario, understand the user needs and psychological characteristics in this scenario, and then in actual operation, once errors are found, adjust the planning and delivery direction as quickly as possible.

In the middle and late stages of a product, the product already has many users and has accumulated a lot of data information. At this time, the user portrait is supported by data. At this time, the data can be used to use algorithms to build models and generate user portraits for the product.

4. How to create user portraits

User portraits can be accurately positioned based on certain data, and the data is divided into two categories: static data and dynamic data.

Static data mainly includes five dimensions: user’s demographic attributes, business attributes, consumption characteristics, lifestyle, and CRM. Within a certain period of time, it will hardly change, such as gender, income, education level, etc., which involve population or other attributes. They are self-contained labels and can generally be obtained directly through user registration.

Dynamic data mainly includes three dimensions: scenarios, media, and paths. When applied to the Internet , scenarios mainly include access devices and access time periods; media refers to the media that users specifically visit in a certain period of time, such as information, video, games , social networking, etc.; paths refer to the paths that users take to enter and leave a certain media, which can be simply understood as the user's on-site and off-site behaviors.

Data aggregation and classification

Basic user classification, such as gender, city distribution, age, income, occupation, etc.

User label annotation (label + weight)

Tags: user interests, preferences, etc.

Weight: user demand, user index, etc.

Time: nodes + length

Location: Content + Browsing Path

Actions: browse, share, order, like, etc.

Perform statistical analysis on the original data to obtain fact labels, and then perform modeling analysis to obtain predicted labels. Finally, from a macro perspective, we summarize and classify the data sources, and use the form of labeling and weighting to obtain the final user portrait.

Of course, the analysis of user portraits needs to be verified, and data is the most reliable basis.

In order to truly understand user behavior and accurately describe user portraits,

I hope you will gain something after reading this article through in-depth exploration and accumulated practical experience.

Author: Junior Sister , authorized to publish by Qinggua Media .

Source: 51coo (ID: COO-51 )