User Response Effect Analysis Basing On DID Model

In modern society,the network is very popular and commodity website has been in everywhere.Gradually,people have pay more attention to the value of the data,and the combination of network and data will also become an unstoppable trend.Competition between commodity website makes the website constantly adopt strategies for effective marketing and attract new customers.Based on this background,we analyze the influence of the campaign on web traffic,and find out the type of potential customers based on the marketing history data.For the effect analysis of customers' response to site,the procedure is divided into three parts.First,the site have released news about new launched campaign through various channels,second part is to collect web pages views data in the period before and after the event,a total of 30 days.Finally,use the collected data to analyze if this campaign is indeed the factor to increase website traffic.To rule out other factors' influence,we use control group and adopt the method of Difference in Difference.The final estimated effect were 388.62(time/day).The test of model shows that effect is significant.In the analysis of customer response to a product,the first is the data analysis and processing,and then build a model,finally is the evaluation on the model performance.In terms of data processing,the data of the sample size and the basic characteristics of the data set is described,and then to improve and upgrade the quality of the data.Using SAS software statement to do the selection of variables and the model fitting,and finally get logistic regression model.The model was combined with dependent variable of customer response and independent variables of 11 variables.According to the modeling results,we can do customer classification on the basis of different response rate,and screening high response rate of customers as the target customers for marketing.This method not only can save cost of mass marketing,but also can improve the ratio of the response.At the end of the paper,the conclusion has carried on the detailed explanation for the model,the application development and shortcomings are also expounded.