from the site of the user level, according to the behavior characteristics of the user access users will be subdivided into various types, because different user behavior, behavior statistics index is different, different view, so if you want to do the user segment, from many angles according to various rules to achieve a variety of different classification, see some of the data analysis report do a variety of user segments, analysis of user behavior, combined with a variety of other dimensions, looks absolutely content rich enough, but it is difficult to understand these results in the end is to explain what the problem, perhaps as an advisory report reflect current trends and the overall characteristics of the user is right, but if you really want to let the data the results of the analysis can guide us what to do, or to determine the purpose of the analysis in customer segmentation, clear industry Business level demand.
wants to do a comparative analysis based on user segmentation, it naturally aims to clarify the difference between the behavior characteristics of certain user classification groups and other user groups. This is mainly from the guidance content level adjustment oriented, by comparing the difference of the user segments on demand content, optimize the contents of operation, quality of content or meet user preferences recommend to the user.
is based on user segmentation, firstly user subdivision rules, for example: here 3 segment loss of users and retained users, new users and old users, single purchase users and two buy users, these 3 kinds of segmentation based on classification, for each user to buy goods in the comparative analysis, to identify the goods more in line with the user’s expectations.
drain users and retained users compared to
, of course, to distinguish the loss of users and retain users, we must first have a clear definition of the loss of users, users can define a loss of active users reference before the blog article, the site and the loss of the user. The definition we can do statistics and subdivision, or to the electronic commerce website as an example, the electricity supplier website content is based on the calculation of each commodity goods, we buy these goods users after purchase may cause the loss of the proportion of users, as follows:
index is defined here should be more clear, the loss of each commodity should be the proportion of users is the number of users to buy the goods after the loss in all users to purchase the commodity in proportion, but the proportion of users only know the loss of each commodity to the evaluation of whether the goods have positive effect on customer retention, or to a certain extent due to the loss of users, only through the comparison to the overall level and draw the corresponding conclusion. So here is the need to focus on the interpretation of "and generally" is how to calculate this value, the percentage is not the result of direct subtraction, but a difference in amplitude reflected, here the assumption that the overall churn rate is 56%, then the A product as an example, and the general comparison is: (58.13%>