Select the remaining 10,000 users in descending

[rifattryo.ut11]

there are 10,000 users in the lost pool and we only have 10,000 SMS budget, how to select these 10,000 people to maximize profits? Solution: Determine the recall benefit coupons. If full-reduction coupons are issued, randomly select the experimental group and the control group. Each user in the experimental group will receive coupons, and the control group will not receive coupons. Collect the results within one conversion cycle, mark the used and unused ones, and build a t-difference model. Use the experimental group and the control group to build a regression binary classification response model.



The marketing users input the model to iran telephone code calculate t. models with targeted users. Use the gain value to prove the value of precision marketing. First, let us introduce the principle of the t model. The gain experiment t is a model used to estimate the individual intervention increment tmtT, that is, the effect of the intervention action ttmt on the user's response behavior.The basic idea of ​​the t-gain experiment is to estimate the response probability of each user under intervention and non-intervention conditions by establishing a prediction model, and then calculate the difference between the two, i.



e., the gain value. How can user operations accurately issue coupons by building a t-model? The principle logic is very simple and crude. It is to predict the difference in the repurchase rate of users under the two situations of intervention and coupon return and non-intervention. This operation can eliminate those users who will return even if coupons are not issued. Secondly, can the operation students build such a prediction model? What kind of person is suitable to be an end product manager? To become an excellent end product manager, understanding business and product are two very important criteria.