User operation: How to build a user portrait model?

At this stage, operators, no matter which position they are in, are required to understand users better. How to understand user needs requires establishing a user model. Establishing a user model includes three aspects: establishing user portraits, conducting user stratification, and analyzing user behavior.

Establishing user portraits is an important part of operational work, so what are user portraits? What is the role of user portraits? How to build a user portrait model?

1. What is User Portrait?

The so-called user portrait, in simple terms, is a labeled user model abstracted based on information such as the user's social attributes, living habits, and consumption behavior. That is to label user information.

For example, Han Meimei, 30 years old, female, married, mother of 2 years old, first-tier city, senior executive

"30 years old", "mother of 2-year-old child", "first-tier city", these are all labels attached to Han Meimei.

2. What is the role of user tags?

What is the use of collecting these user tags? Based on these tags, we can understand user preferences, explore user needs, conduct targeted marketing, make personalized recommendations, and improve the conversion rate of advertising.

For operations, establishing user portraits can greatly improve operational efficiency and help refine operations.

3. How to build user portraits for products in different life cycles?

The role of user portraits is different for products in different life cycles. For example, for start-up products, due to their insufficient number of users, user portraits are just a general concept and the company is often not very aware of the need to build user portraits. Building user portraits is just to understand the target users more accurately.

Therefore, the methods and channels for products in different life cycles are different.

Regardless of the stage of the product, to build a user portrait, we can start by collecting user information data.

1. User Information Label

The collection of user information includes user basic elements, user scenarios, behavioral preferences, psychological personality, communication and other aspects. It can be divided into two categories: static tags and dynamic tags.

It is worth noting that user information data covers a very wide range, especially dynamic tag data is variable, so establishing a user portrait model is also a process of continuous improvement.

User Information Tags

The above are some of the tags listed for user information data.

Static information tags are the basic framework of user portraits. In operational work, we can refer to the static information models of mature enterprises, which can help us improve the portrait dimensions as quickly as possible.

But in operational work, dynamic information can often provide us with more useful key information.

The so-called dynamic information tags actually refer to the user's online behavior. It includes a series of behaviors such as searching, browsing, registering, logging in, signing in, posting information, collecting, commenting, liking, sharing, adding to shopping cart, purchasing, using coupons, using points, etc.

User behavior depends on the attributes of the website. For example, e-commerce websites can collect more user consumption attributes. Social networking sites make it easier to collect users’ social attributes.

In addition to user behavior, it also includes dynamic information such as the frequency of user behavior, duration of user behavior, time period, and device.

By counting real user behaviors, users are labeled with different behavior labels, and then model labels are established. For example, demographic attributes, user activity, user interests and hobbies, user satisfaction, channel preference usage, purchasing preference, content preference, user associations, user risk scores, etc.

Through model labels, further prediction labels are established, such as: population attributes, consumption capacity, churn probability, default probability, recent demand, and potential demand.

Take WeChat public account as an example:

Static data includes: user's gender, age, occupation, work city, graduation school, etc.

Dynamic data includes: following a public account, reading an article at a certain time, liking an article, leaving a message on an article, sharing an article, the device terminal used, unfollowing, etc.

Further dynamic data include: how long you stay on an article, the reading completion rate, etc.

In addition to studying users' factual data, studying users' psychological phenomena, especially the three aspects of needs, motivations, and values , can help us gain insight into users' deep motivations for registering, using, and purchasing products; understand users' needs for product functions and services; and recognize what value labels target users have and what kind of group they are. However, this is a more complex research direction.

2. How to collect user information data for products with different life cycles

(1) Start-up products: User profiling during this period is more about collecting static tags, identifying target users, and optimizing and improving product functions, design, content, etc. Understand the attributes of target users and build traffic.

At this stage, the product itself has less user information data, so a general survey of user information can be conducted by investigating competitors and conducting direct user questionnaires.

(2) Growth stage products: The user portrait at this stage is important for precision marketing, which makes the product's service objects more focused and dedicated, better meets user needs, optimizes operational methods, and improves the company's operating efficiency.

Products at this stage have been recognized by the market, and various data are on the rise. You can directly organize the data fed back by the product background to create a more detailed user portrait.

(3) Mature products: Products at this stage have basically stabilized their market position, and their daily work is mostly maintenance-oriented. We need to find new growth points and breakthroughs through user portraits.

The user portrait at this stage can be made by combining the general survey in the start-up period with the user portraits accumulated during the growth period.

Summary: The above are some suggestions on the dimensions and channels for building user portraits. It is worth noting that user information covers a very wide range, especially dynamic tag data is variable, so establishing a user portrait model is also a process of continuous improvement.

Source: Random Thoughts