Application Of The Linear Mixed Model In Marketing Mix Optimization

Marketing is the core of enterprise management and operation.The development of network technology has made some digital new media become the powerful tools for enterprise marketing.Companies must develop a reasonable and effective marketing mix strategy,in order to enhance sales performance,as far as possible to obtain maximum profits.First we has carried on the statistical modeling of advertising and marketing data of the theme park over a period,and then looked for the optimal marketing mix based the model.In addition,with the improvement of people’s living standard,more and more people pursue the enjoyment of spiritual life.One of the highlights is the rapid development of tourism industry,such as all kinds of the theme parks around the world.These companies’operation mode is online and offline are integrated tightly,and their key target customer group is young people,so all kinds of media advertising and marketing are particularly important for this kind of company.Our advertising and marketing data is derived from a theme park,covering every week from 2012 to 2015 and 210 designated marketing areas(DMA)of the US.The total observation amount is 42010.Variables include different advertising media channels measure,and ticket sales variables,etc.Because our data is area and time level observations,so we first focus on the integration of original data of advertising channels,then carried on the descriptive statistical analysis,calculated numerical values such as the mean value and standard of variables.Also we depicted the time line chart of the marketing channel variables to analyze their trend,cycle time,etc.The mixed effect model is established in order to achieve a better fitting effect.Considering the effect of advertising decay,lag and saturation of the actual effect,we transformed advertising channel variables,and then put into the model for variable selection and the determination of deformation parameters.According to the results of the model,media contribution rate and return on investment of last cycle’s marketing activities,we push the process of variable selection until the final form of model and model results are determined.In order to find the best marketing mix,we got from the fitting model to calculate the amount of the contribution of each marketing channel and the return on investment(ROI).Total amount of contribution equals to the regression coefficient β times weekly input of each variable,and then aggregate week results.ROI is made up of media channels sales divided by the cost of each channel input.The results show that the higher investment yield,marketing media are digital display platform,direct mail and paid search.Therefore,according to the research of this paper,the next cycle marketing activities of the theme park should value the role of digital new media,move the center of gravity to those channels,and optimize the marketing mix,to achieve maximum profits.