User operation: 8 tricks for user portrait model!

User portraits are a way of describing and depicting user characteristics in a labeled form. As a basic tool for comprehensive insight into users, user portraits are widely used in precision marketing, personalized services and other fields.

However, creating user portraits is similar to being a product manager: it is easy to get started but difficult to do well. It is no secret that the user models of many companies are just fancy words, bland and empty, superficial and ineffective, and they always feel a little lacking in use.

The reason for this is that the product manager of the user portrait did not think deeply from the user's perspective and did not fully understand the needs of business personnel in using the portrait. So, the question is, how can we make user portraits produce better results? The author has summarized eight tricks, hoping that they will be of some inspiration to everyone.

1. Combining static and dynamic elements to make user portraits more three-dimensional and vivid

If there are mostly static tags, the analysis results of the user portrait will be superficial and lack depth, and the output will be some charts such as gender, age, occupation, education level, etc. These are all possible, but they are obviously not enough.

The author believes that the label combination of user portraits should be a combination of static and dynamic, with both static and dynamic labels. Static labels show relatively stable characteristics, while dynamic labels show nearly real-time characteristics. Through the flexible combination of dynamic and static labels, the user's image can be made more three-dimensional and vivid.

For example: After collecting data on users’ terminal information, tariff packages, traffic usage behavior, etc., the customer service center of a mobile company drew a user portrait in the form of a combination of "dynamic tags + static tags".

Static tags describe information that will not change in the short term, such as user value, terminal model, basic tariff, etc., while dynamic tags describe a series of actions taken by the user in the last month - from receiving SMS reminders to calling 10086 for consultation to handling business at the SMS business hall. When regularly scanning existing users on the network and looking for target users who meet the above profile characteristics, you can actively recommend 5 yuan traffic packages.

Facts have proved that this user portrait that combines dynamic and static elements is more accurate than simply relying on static tags to filter target users.

2. Coarse and fine combination to make user portraits closer to application scenarios

The labels selected for user portraits do not necessarily have to be at the same granularity. Some labels can be coarse-grained, while others can be at the finest granularity. This depends on the combination of specific business scenario requirements.

A basic principle is that the more relevant a tag is to a business scenario, the finer its granularity should be, while the less relevant a tag is to a business scenario, the more coarse its granularity can be. The more important the tag is, the finer its granularity should be, but it can be as coarse as possible.

In short, we need to combine the rough and the fine to make the user portrait more closely related to the actual application scenario.

3. See the big picture from the small details and find valuable labels from the details

Previously in the TV series "The Inspector General of the Song Dynasty", we often saw Song Ci analyzing clues at the crime scene to infer the identity characteristics of the suspect, and give an accurate portrait of the suspect. He would always find the real criminal by following the portrait of the suspect.

What I want to emphasize here is that a true portrait master can see the big picture from the details, discover some valuable information from the details, and then make a clearer portrait of the target person based on the combination and processing of this information.

In the business field, when companies are creating user portraits, if they can collect the user's QQ avatar and signature, WeChat avatar and username, handwriting, etc., this information is actually related to the user's personality and psychological characteristics.

Studies have shown that QQ avatars reflect the user's preferences and personality. For example, people who use photos of themselves and their families as avatars tend to be self-centered and stubborn; people who use cartoon images as avatars tend to be nostalgic and idealistic.

For example, studies have shown that handwriting can also reflect a person's personality traits, and handwriting can be used as evidence for personal identity verification. Of course, nothing is absolute in the world, and this is just a general trend. However, information such as QQ avatars and handwriting can indirectly reflect the personality and psychological characteristics of some subjects, and there is definitely room for application in user portraits in the business field.

4. From the outside to the inside, introduce labels on personality, psychology, values, etc.

As the saying goes, “It is easy to draw the skin of a tiger but difficult to draw its bones” and “You can know a person’s face but not his heart”. User portraits can generally describe the user’s representative information, but they rarely depict deep-level labels such as the user’s psychology, personality, and values.

In the suspense TV series "Psychological Crime", Fang Mu (played by Li Yifeng) has some wonderful classic operations of character psychological portrait. Through research and analysis of clues at the crime scene and the murderer's modus operandi, Fang Mu was able to infer the murderer's psychology, past experiences, etc., and give a reasonable portrait of the murderer, laying the foundation for the subsequent rapid resolution of the case.

The reason why we should pay attention to deep-level labels such as personality, psychology, and values ​​is because these are the most basic determining factors that make people produce corresponding behavioral characteristics. In other words, the "inside" determines the "face", and the most basic labels determine the surface user characteristic labels.

By the same token, if companies can incorporate some user consumer psychology evaluation techniques and methods into their business activities, and pay more attention to deeper information such as the user's personality, emotions, psychology and values, it will undoubtedly help make more reasonable business decisions based on a more accurate grasp of user characteristics.

5. Use flowers to make the label more vivid and vivid

When labeling users, in order to make the labels more distinctive and recognizable, and easier for users to remember, we can learn from others and incorporate more classic character symbols as labels. For example: a person who is lazy and likes to play tricks can be labeled as "Zhu Bajie", and a person who is resourceful and foresighted can be labeled as "Zhuge Kongming".

The reason why these classic characters can be used to label target users is that these characters are recognized by the general public, their images and characteristics have been deeply rooted in people's hearts, and they themselves have strong explanatory power. Labels like "Zhu Bajie" actually concretize and nickname the target user's "personality" and label, making the user's label more vivid and vivid, and making the user portrait easier for users to understand and use.

6. Go deep into the front line and listen to the opinions of front-line business personnel on user portraits

The people who understand users best are often not product managers, but front-line business staff, because front-line business staff are always dealing with users and understand users' psychology and characteristics best.

Therefore, when making user portraits, you should listen more to the opinions of front-line business personnel and let them describe the user portraits in their minds. Product managers need to translate their descriptions into labeled data language.

Of course, product managers cannot just listen to front-line business personnel. They still need to incorporate their own understanding and thinking so as to extract a down-to-earth user portrait.

7. Keep improving and find the North Star label combination

User tags have a life cycle, and the tags that are suitable for a user will change at different stages of his or her life.

In actual applications, user tags need to be tested in various activities, and user portraits also need to be iterated and improved in different business practices. User labels and their combinations need to be constantly refined, and machine learning should be used to make labels more accurate and label weights more reasonable, ultimately forming a label combination similar to the North Star.

The product manager of user portraits must strive for excellence and find the North Star label combination for users in different scenarios.

8. Application-oriented, organize user portrait content based on business scenarios

User portraits cannot just be for your own enjoyment, they need to serve the business and application scenarios. The author firmly believes that a good user model should start from the end in mind, be application-oriented, serve the users of the portrait, and constrain and organize the portrait model based on the purpose of the application.

Therefore, the content displayed by the user portrait should echo the user's needs and should have a high relevance to the application scenario. Only in this way can the user portrait be close to the user's needs and play its due value.

summary

The correct way to open a user portrait is to be application-oriented and combined with business scenarios. User tags should not pursue quantity but emphasize quality. Customized tag combinations are required in different scenarios. Based on this principle and making good use of the eight tips I mentioned earlier, the user portraits you are responsible for will make you the most popular person in the company.

Finally, do you want to defend the product value of user portraits? So, please make use of the above eight tips to make your user portrait more effective!

Author: Big Data Product Design and Operation

Source: Big Data Product Design and Operation (gh_a9d6d925b696)

Related reading:

From 0 to 1, how to build a user portrait system?

User operation: How to build high-quality user portraits?

User Operations: Are you sure you really understand user portraits?

User operation: How to build a user portrait model?

User Operation: How to create a To B user portrait?