RapidMMM: Richer, faster and affordable marketing mix models

Patralekha Bhattachary and Krishna Mehta
Jigyasa Analytics


Marketers have always struggled to accurately measure the impact of their advertising campaigns on Sales. This task has become more challenging today because of the numerous types of different marketing activities that are available to the media planner. This coupled with limited and sometimes incomplete data makes estimating marketing impacts a lengthy and expensive process. The time and costs involved in building a marketing mix model implies that model rebuild is an infrequent phenomenon and often the performance of these models deteriorate significantly before they are evaluated and improved. This limits the ability of firms to react quickly to changing market conditions. In this paper we introduce our proprietary RapidMMM solution which uses advanced statistical and machine learning techniques to build marketing mix models within hours. This improves the richness of the models as well as significantly reduces the time and costs involved with building a marketing mix model, making these models affordable to a larger cohort of media planners.