User identity identification and account system practice

The internet account has its own memo mechanism ;

One 、 Business background

Usually in the process of system development , We need to constantly adapt to various business scenarios , Expand service areas and capabilities , Generally, the product matrix will be divided into multiple business lines , In order to better manage ;

Due to the different data entry and management strategies of each business line , So the data precipitated from different paths , It may be isolated because of the boundary problem of the system ; If user data is split , It will mislead the analysis and decision-making because the data is not comprehensive ;

A classic scene , After the user completes the registration from the application , Usually don't provide too much information about yourself , Because the business needs to constantly enrich user portraits , So user data is usually dispatched to an independent management system , Information expansion through different contact feedback , Such as collecting buried point data , Offline contact , Marketing calls, etc ;

This situation is a scene with obvious perception from the operation , Obviously, there is a big difference between users' data in the application library and that in the management Library , In the real situation, users may have repetition in different applications and scenarios , It will inevitably lead to difficulties in unified maintenance of user data ;

Two 、 Unique identification

User behavior data is in current Internet products , It is of great analytical value , Different application terminals are logged in or not , The data generated in the product is a recorded means , Then analyze and identify at the data level ;

The biggest feature of these numbers is their uniqueness , It can identify the operation information of users in different terminals and different states , And when these data precipitate into the system , It will be stored according to the port and operation type , The unique identification of data is also different under different terminals ;

From the perspective of data analysis , Obviously, we don't want users' behavior information to be split and isolated , In this way, the user behavior data in multi terminal and multi state can be correlated in the whole domain , Is an effective way , Its basic principle involves ID Mapping technology of ;

3、 ... and 、Id mapping

Based on the above business situation , It is very important to provide a global unified identification of user identity in the product matrix , Behavior data generated by user entities in different business lines , Identification by unique serial number , In this way, the image seen during user analysis is more comprehensive ;

In the current Internet products , The mode of creating application account based on mobile phone number has been a common function , After the mobile number is registered , Then associate the corresponding terminal through the mobile phone number ID, So that all kinds of isolated data are linked ;

The principle of its implementation is not complicated , First, you need to provide a set of mapping libraries , When the new mobile phone number is recognized and collected by the system , Create a new piece of data in the mapping Library , Mobile phone number and corresponding unique ID, After that, the data of other paths , If the mobile phone number is the same, it will be bound in this ID below ;

Four 、 Data Association

stay ID Under the mapping mechanism , Although the data of each business line is relatively isolated , There will be no direct impact between data , But in fact, it has been the only ID In series , This will ID Comprehensive analysis of the related data , Accuracy will improve a lot ;

No matter the data collected from any path or channel , If there is a dimension of mobile number , Or the serial number identification associated with the mobile phone number , Judge whether there is a global mapping for the mobile phone number ID, If not, create a corresponding relationship in the mapping Library , If yes, bind directly ;

When performing global scheduling and analysis of data , Through the standard relationship of the mapping Library , be based on ID Identify the data of all business lines for query and overall analysis , So as to generate a relatively comprehensive data file , And standard analytical logic ; Here is a reference structural design ：

There is logic of Data Association ,ID Logo and mobile number are unique and one-to-one , But there may be one to many between the mobile phone number and the serial number of the terminal , Even many to many ; Account number and behavior data generated in the application , Although the pursuit of accuracy , But accuracy is not overly required ;

In this case, it is necessary to implement the corresponding business strategy , For example, the same mobile number may have logged in to the same application in different mobile phones , Applications in mobile phones may also have been logged in by multiple accounts , At this time, we need to make trade-offs based on policies , It may be the login duration of the account , Or the time period before and after login , It's impossible to generalize ;

5、 ... and 、 Register login

Take the mobile phone number as the account subject , Open applications don't make a big difference between registering and logging in , In order to simplify the operation and avoid blocking the user , When logging in through your mobile number , If you are an unregistered user, you can initialize the information directly ;

• The user is in the login form , Enter the phone number and get the verification code ;

• In the login service , Generate and maintain the timeliness of verification code ;

• The verification code needs to be pushed to the user's mobile phone with the help of the docked third-party SMS platform ;

• After filling in the verification code in the login form, submit the login information for verification ;

• After login verification is successful , If the user is not registered, initialize the account system ;

• After account system verification and maintenance , Relate asynchronously ID identification ;

• Finally, you need to return to the client Token Identity token , As an account identification ;

The reuse interface integrated with registration and login is relatively complex , But let users use the product quickly with the shortest path , Through behavioral data collection and analysis , Thus, user needs can be accurately identified , Conduct correct guidance and marketing , Give play to the real value of data ;

Here is a reference for the structural design of account management , Usually, the user's main table dimension will be designed around the login account , The data related to information collection will be written into the user profile table , Because different business scenarios depend on different information , Therefore, after user registration, various data collection pages will be guided ;

User identification and account number are the core capabilities of the system , When designing, we should have user experience , At the same time, we should pay attention to the security of data ; As a core ability, we need to be forward-looking in the early design , Plan the possibility and reserve the structure , Avoid subsequent iterations that span too much .

END