Website user's identification grasps the user demand to enhance the user experience

Intermediary transaction SEO diagnosis Taobao guest Cloud host technology Hall

User analysis is an important part of the analysis of the Web site, before the analysis of users we must first be able to identify each user, identify which is "New customer", which is "Repeat customer." This will not only provide a clearer picture of how many users have visited your site, identify who they are (user ID, mailbox, Gender age, etc.) and also can help you to better track your users, find their behavior characteristics, hobbies and personalized settings, so as to better grasp the needs of users, improve the user experience.

Usually when your site provides a registration service, and users sign up and log on to your site, then users can be more easily identified, because the site will generally save the details of the registered user, but your site does not need to register, and the user's behavior to browse the main, which is user identification will appear more difficult, Several common methods of user identification are provided below:

Several ways to identify users

When the user does not register to sign in, the only way to identify the user is only the user browsing behavior of the click Stream data, usually they will be saved in the Web log, the detailed description of the Web log can refer to my previous Article Web log format. And the Web log itself defects may lead to inaccurate user identification, about the Web log defects can refer to the previous Article Web log function and defects, so we select the user identification method in the process, when the conditions allow to choose a more accurate method:

1, ip-based user identification

IP address is the most easily obtained information, any Web log will contain, but its limitations are also more obvious: pseudo-IP, agent, dynamic IP, LAN share the same public network IP export ... These situations will affect the IP to identify the accuracy of users, so the accuracy of IP recognition users is relatively low, the current general will not directly use IP to identify users.

Get Difficulty: ★

Accuracy: ★

2. User identification based on Ip+agent

Also based on the simplest form of web logs, we can add a--agent to improve the accuracy of a single IP to identify the user. The agent is also a Web log in general will contain information, through the ip+agent way can appropriately improve IP proxy, public IP, such as the user's resolution, while the agent can also identify the network crawler special "users", but the same accuracy is not high.

Get Difficulty: ★

Accuracy: ★

3. User identification based on cookies

When you get user cookies by customizing the Apache log format or JavaScript, you've actually found a more effective way to identify users. Cookies can be considered to be bound to an Access client computer (a client may contain multiple cookies) without being purged, so using cookies to identify the user is actually referring to the client computer that the user is using, not the user itself.

The method of identifying users with cookies is of course flawed: the most common is that cookies are purged, causing the user to not correspond to the original record, and because the client computer is shared, or if the user accesses your site on a different computer, the cookie cannot be directly mapped to the user.

Getting difficulty: ★☆

Accuracy: ★★☆

4. User identification based on user ID

User identification based on user ID is the most accurate, because in general the user is different share his user ID, so we can think that the data in the UserID unique point to the user, almost no deviation. Of course, to use the user ID to identify the user is required a certain prerequisite: the site must be provided by the user Registration login service, and can be a number of means in the click Stream Data recorded UserID.

Get Difficulty: ★

Accuracy: ★★★

So for a Web site that requires a user ID registration, the user's unique identifier can be chosen in the following order: When the user registers the login with the UserID, when the user is not logged in state browsing to the user's cookie, when the user is not logged in and the cookie is not available to ip+ The agent is the same, so that you can identify the unique user.

A custom setting method for cookie items in a Web site log is recommended to better identify users. Cookies are obtained from the cookie file record stored in the client, which typically contains a cookieid and also notes the user's UserID on the site (if your site needs to be registered and the user has logged on to your site and the cookie has not been deleted), So in the log file cookies can be a priority to query whether the cookie contains the user ID class information, if there is a user ID written to the log of the cookie item, if it does not exist to find whether there are cookieid, if there are records, no is recorded as "-", This allows the cookies in the log to be used directly as statistics for the most efficient user-unique identifier. Of course, it is necessary to note that this method only can be implemented by the site itself, because the user ID as the user's privacy information only the site to know its cookies in the settings and storage location, third-party Statistical tools are generally difficult to obtain.

Ways to get user information

Through the above methods to realize the unique identity of the user, we can through some channels to collect the user's basic information, characteristics and behavior information, and then for each user to establish a detailed profile:

1 user registration information and basic information;

2 The user browsing behavior data from the website log;

3 The user website business application data obtained from the database;

4 The derivation and prediction based on the user's historical data;

5 The user data obtained through direct contact with the user or user survey;

6 The user data provided by the Third party service organization.

Identify and capture the value of user information

Through the user identification and the user basic information collection, we may through the website analysis various methods in the website is realizes some valuable application:

User segmentation based on user characteristic information;

Personalized page settings based on the user;

Association recommendation based on user behavior data;

Oriented marketing based on user interest;

......

» This article uses the»in agreement, reprint please indicate the Source: Website Data analysis» "website user's identification"