Usually, we will also search the user's portrait on the Internet. Generally, we will find two answers.
The first one is to tell you that users prefer the content of the tag number.
The second is to describe a user's professional interests and some characteristics when using products.
There will also be some gaps between the two. The first generally refers to some behaviors and characteristics of users in the system, or the information voluntarily filled in by users, which is more inclined to the label system formed in the process of data statistics. The second kind is more about partial users interviewing users, some emotional things.
The first one will be used in product promotion and operation and product design. This kind of performance usually comes from the specific data of products, so it will be used more in business or operation. The second kind will generally help us identify the types of people that products mainly serve, and it is a more emotional thing. But it will be missing when guiding our specific work.
There are some differences between the two, but we need to know more about the first user portrait based on data.
Basic attributes (user basic information): age, gender, birthday, constellation, education, height, income, occupation? ....
Social relations: marriage, children, girls, the elderly at home, sexual orientation. ....
Behavior characteristics: basic behavior registration time, source channel ...? Business behavior bought preferential goods and won excellent students.
Business-related (sports products): tall and thin, body fat rate, muscle exercise, 9000 steps per day, collection 100 fitness programs. ....
1, fill in directly (for example, when registering, blind date products, take-away products and decoration products).
2. Originated from users' own existing characteristics (generally, when doing activities, the operation is simple and personalized, and business analysis should be observed in groups and user research (preparation))
For example, if the product is an e-commerce platform, the operation department will hold an operation activity such as Girls' Day with female college students in Beijing.
At this time, how to distinguish gender, address and spending power?
We can use the existing functions to push. We can infer from what we have bought, for example, those who have bought many male products are classified as men, and those who have bought many kinds of sanitary towel products are classified as women. The address can also be inferred from the receiving address. Consumption ability can also be inferred from consumption details (such as buying a consumable with more than 200 yuan) and the characteristics of users, which are screened out through previous information. It is impossible for all users to be like this. Only some users can be labeled and some users don't recognize it.
Of course, we need to make further derivation. For example, through the user's common IP, you can further find out whether you are in Beijing. Or spending power can also check whether the mobile phone model used by the user is new. For example, through the mobile phone models vivo and Mito mobile phones, it can be concluded that most female users (of course, there will be some misjudgments) can use the receiving address, university consumption and other places.
3. Infer from the people around the user.
Close distance (some attributes are owned by people around, and the probability of users is high)
Similar behaviors (target users with similar behaviors are found through collaborative filtering)
Common characteristics: basic attributes, social relations, behavioral characteristics and business relevance.
Usage scenarios: marketing, personalized operation, business analysis and user research. ....
There are generally three landing scenarios (high-quality innovation, precise operation and push, and auxiliary product design) to understand the users behind the numbers through user portraits.
How to pull high-quality new products? (second-hand book trading platform)
1 How to find us from existing users, who are the real users?
Define what our real users are (such as high retention users, core behavior frequency, high completion rate).
2 characteristics of real users
who is it? The e-commerce platform can pass the books he bought. Turn their age, education level, region and spending power in reverse.
Where did it come from? ? Telephone interview, etc. Many of them were recommended by friends.
3 Find similar users by this type.
? Portrait of users: universities, research institutes and knowledge-intensive work fields. Consumption tendency of social science books.
Sometimes this method of pulling people may not be done well. There will be many conditions that limit us and there is no way to do it. In this case, we can choose many channels to cooperate. Different channels will have different attributes, and users of different channels will have different labels, such as age labels, professional gender and so on. With your own label, you can choose a better delivery channel according to your own demands and characteristics when docking channels.
In addition, there are many well-done advertisements, all of which can accurately select the audience. Comparing the labels and portraits of the two platforms can be correct. But at the beginning, we must figure out what kind of people we want to put in, otherwise even the best platform will not be used. After we have the user portrait, we will have a general reference when we pull the new one.