Site user identification, grasp the user needs, enhance user experience

analysis is an important part of the site analysis, in the analysis of the user before we must first be able to identify each user, identify what is "New Customer", which is "Repeat Customer". This will not only be able to more clearly understand exactly how many users visit your site, identify who they are (user ID, email, sex and age etc.); at the same time also can help you better keep track of your users, found the behavior characteristics, their hobbies and personalized settings, in order to better grasp the user needs, enhance the user experience.

usually when your site provides registration services, and user registration and login your website, so users can more easily be identified because details of the website will generally save the registered users of the site; but you do not need to register, and the behavior of users to browse, the user identification will appear the more difficult, the following provides several methods of user identification:

Several ways to identify users by

when the user did not sign up case, the only way to identify the user is the only user browsing behavior of click stream data, usually they will be saved in the WEB log, a detailed explanation of the WEB log can refer to my previous article: WEB log format. WEB log defects itself may lead to the inaccuracy of user identification, defect on the WEB log can the function and defects of WEB log reference before, so we in the choice of user identification method, under the conditions allowed to choose a more accurate method:

1, IP based user identification


IP address is the easiest to access information that will be included in any WEB log, but its limitations are obvious: pseudo IP, proxy, dynamic IP, LAN sharing, and the same public IP exit…… These situations can affect the accuracy of user identification based on IP, so the accuracy of IP identification is relatively low, and IP is not directly used to identify users.

: the difficulty in obtaining.


2, IP+Agent based user identification

, too, based on the simplest form of WEB logs, we can add an item — Agent, to improve the accuracy of user identification in a single IP mode. Agent is the WEB log will generally contain information, by way of IP+Agent can improve the IP agent, the resolution of this kind of situation of public IP users, at the same time through the Agent web crawler can also identify the special "user", but also the accuracy is not high.

: the difficulty in obtaining.

: accuracy.

3, based on user identification cookie