Automated analysis and planning of social network marketing

Abstract

Managing online marketing campaigns is a repetitive and analytical process which is typically done manually by domain experts. This paper deals with the problem of how to use software to manage historical marketing data and use that as a foundation for decision-making with the purpose of optimizing future ad performance. The solution presented in this thesis involves applying both data mining and automation practices to provide decision makers with knowledge that can be enacted upon. Operators of the system input the historical data, upon which the system creates an estimation model based on said data which is used to estimate the performance of, for the system, previously unseen ads. These suggested ads can either be input by the operator or automatically created by combining existing ad properties. The solution has been validated on real-world data from an industrial partner in the social networking industry.