On the so-called product "User Portrait"

In the first half of 2015, China's Internet users have reached 668 million, the end of the year will certainly be able to break through 700 million, which use mobile phone online population accounted for 88.9%. Different from the traditional PC Internet, each family share a device, mobile Internet has a unique, unique and private characteristics, each person's mobile phone is a unique set of ecological system. Therefore, users with the same characteristics are abstracted into a representative, which can greatly facilitate developers to study user composition and distribution, and define users precisely. This is also the so-called Big data, precision marketing, and other concepts of explosive fire reasons.

China in all respects is a big long tail market, the Internet to a large extent make up for information asymmetry, mobile Internet and let the information in the accurate delivery to any user, regardless of time and place. So the question is, how can we realize the business value of the product by realizing the flow of traffic?

To give full play to the true value of big data, the first step should be to collate the data. And the stage of organizing the data is to complete the target user or the "portrait" of the existing user, only to get the accurate user portrait, can better achieve the purpose of traffic monetization. 1. What is "User portrait"?

Users, refers to your target users, or the majority of the groups that make up the existing users collectively. A portrait is an objective, accurate, visual description of a thing. Then "User Portrait" is the ability to objectively, accurately, visually describe the target user's tools or methods . 2. How to create a "user portrait"

A good "user portrait" has two important principles, one is the label, the second is the low crossover rate. Labeling of

The so-called labeling, refers to the product needs, to different user characteristics affixed to the appropriate label.

The dimensions of the label need to be divided by product and business needs, such as the "purchasing power" dimension in e-commerce products, educational products, pure tool product user portrait of the importance of the degree is completely different, it may be in the e-commerce products in addition to the basic characteristics of the most important indicators, but in the pure tool products may not need statistics.

If the size of the image needs to be refined a bit more, you can further add weight to the label. For example also hit the "High purchasing power" label users, can be divided into such as luxury consumption, electronic products consumption, household goods consumption, and so on, and then according to the target user attributes and product business development direction to the user labeled after the subdivision, while the "purchasing power high" allocation of the corresponding weight.

The goal of labeling is to abstract user attributes in an electronic way, to facilitate data-counting colleagues to build data marts, and to follow up on data mining and aggregation analysis.

After the effective data mining classification results is what we really need, as to what the process is, we can adjust according to the actual situation. But do not shoot the forehead to label users, such as the general characteristics of students are blind, limited consumption capacity, clothing to sports brand-oriented, like the new strange things, our products happen to be new strange things, so students are our target group. This conclusion is not a problem to put Buddha, but the actual student groups like things are group will, not blind, and different stages of student spending span is very large, and so on. Therefore must according to the actual data to classify the user, should not clap forehead. Low crossover rate

The purpose of the "User portrait" is to digitally aggregate and describe the user in the real world using tags.

When aggregating portraits, it is important to pay attention to the principle of low crossover rate. If the two groups of portraits in addition to education and age of the rest of the label is almost consistent, it can be initially understood as the target users, different educational level or different age span of users on the product demand is quite consistent, if there is no obvious business distinction, can be combined into a portrait, while weakening the difference between education and age.

Just now the reverse expression of this principle, the positive expression is the user image screening as far as possible to comply with the Mece law, that is, integrity and independence.

Integrity to ensure that our "User Portrait" includes the largest possible user groups, independence and ensure that the group does not overlap, to prevent the blind exaggeration of the market structure, influence on the market judgment. These two points for us to analyze the market, optimize the product business has a great effect. 3. Dimensions of the user tag

Although it is not possible to shoot the forehead to define the target user, we need to define the label dimension based on the product characteristics. At this time, attention must be objective, do not have to "get XX data, can prove xxxx" mentality , this mentality is unable to establish the correct user portrait. " Typically, a label that creates a user portrait has the following dimensions: Natural features or Basic properties:

such as gender, age, body shape, region, occupation, constellation, educational level, etc. consumption characteristics or purchasing ability:

such as marriage, income, car, room, children, shopping type, brand preference, credit level, purchase cycle, etc. social characteristics or behavioral characteristics:

such as marital status, family composition, social preferences, information channels, etc.; interest characteristics or psychological characteristics:

such as hobbies, use of app behavior, browse collection content, interactive content, etc.;

There is no clear requirement for the dimensions of the label, and even the above classification has no definite boundaries. How to choose the right label to describe the user's characteristics is the question we need to really think about. 4."User Portrait" Final destination

Having said so much about how to create "user portraits", how do you use this when it's built?

Different people have different statements, some think it should be made into the report archive, some think to draw a chart to facilitate the qualitative analysis, some think that the name and introduced to the team members. I personally recommend the last way to give each "user portrait" a proper name and allow everyone on the team to accurately say their tags. Only in this way, the team members can strengthen the target users, but also to facilitate communication problems, consider the need to directly say "in order to meet the needs of Xiao Ming" on it, direct and powerful. 5. Summary

"User Portrait" is not omnipotent, is not essential, but we can be clear that the right and reasonable "user portrait" will be able to promote the development of products, it is a strong proof of the next step in the direction of product development.

So how do you get a user tag? The most common way is to count and analyze user behavior paths. As for the different products how to get data, how to management, product sample small, feature extraction is not obvious problems can only be handed out to each team to solve. Otherwise, why is big data so valuable? This also gives the "User portrait" blind worship of people to wake up, if there are small companies, small team and you said he can rely on "big data" to help you build "User Portrait", first think about his platform is not big enough, enough data, not to be fooled by the concept.

PS 1: To be honest, I've never had a "user portrait", just a recent business need so study a bit, said the wrong place hope you correct;

PS 2: Distressed on December 25, 2015 see this article of people, because you and I like Christmas no one accompany, t_t;

PS 3: I understand the "user portrait" is the following image of the sauce, but the feeling has been 10,000 damage, heart pain.

"User Portrait" that hurts 10,000 points