The data gain of Y will drop significantly, that is, in economics A substantial bulk sms service reduction in marginal revenue, that point is the "inflection point" in the graph. For example, when the churn period increases to 5 weeks in the figure, the reduction rate of the user return visit rate drops significantly, so 5 weeks here is the inflection point, we can use 5 weeks as the period to define user churn, that is, a user who has visited/logged in before , if there is no bulk sms service access/login for 5 consecutive weeks, the user is defined as churn. The second is to count the time interval between the user's last visit and the penultimate visit. It can be considered that users longer than this time interval will basically not visit again, that is, users have been lost. Looking at the historical data,
It can be seen that less than of users have an bulk sms service interval of more than 30 days between the last visit and the penultimate visit, and users who are longer than this time interval can be considered as "churned users". After the churn cycle is divided, users are further divided into high, medium and low activity according to their active status. The historical data is divided according to the 28 principle. For example, the analysis shows that users who have been bulk sms service active for more than 10 times account for of the number of users who visit in the past 30 days, then this batch is "highly active users"; further, users who have been active for 5-10 times are classified as "moderately active users"; Users who are active 1-5 times are classified as "low active users". (2) RFM label .
The RFM model is mainly composed of three basic bulk sms service indicators: the last consumption time, consumption frequency, and consumption amount. According to the historical data, check the proportion of users, which can be divided according to the principle of 28, and the subdivision labels can be obtained. R: For example, in the historical data, of the users who visited recently <60 days are "near", and users who have recently visited >= 60 days bulk sms service are "far". F: If 8 of the historical transaction order volume is the user's order volume < 5 orders is "low frequency", and the order volume >= 5 orders is "high frequency". M: For example, if of the historical transaction order amount is less than 500 yuan, the user's transaction amount is "low", and the