Asset Publisher


Predictive Analytics: Marketing Mix Model

Marketing Mix Model Project Delivers Tool for Marketing Spend and Allocation

Solution: Analytics, Predictive Analytics, Marketing Mix Modeling, Machine Learning

Industry: Food & Beverage

Project: Marketing Mix Model Project Delivers Tool for Marketing Spend and Allocation
Client: US Fast Food Chain

Our client was a U.S. fast-food chain which began over 80 years ago and today has over 36,000 restaurants in over 100 nations. The company has worked hard to be socially responsible and continue to innovate operations. Their continued growth strategy includes a $19MM marketing budget, covering a wide range of offline and online media channels, couponing, and point-of-purchase displays. 

Challenge: Marketing Budget Was Not Optimized

The company advertised in many geographies through multiple media channels.  Because of the extensive reach and channel mix, there was no method of measuring how effective the budget was affecting sales. The client challenged Dunn Solutions to measure the effectiveness of the $19MM marketing budget to measure the Return on Marketing Investment for every media channel and geography. In addition, we were tasked with creating a tool that could be used to plan the marketing budget for future campaigns.  

Solution: Dunn Solutions Creates a Marketing Mix Model to Measure Marketing Performance

Working with the clients’ marketing agency and IT department, we collected two years of weekly sales data spanning 8 geographies and 5 different product categories, along with marketing spend and in-store tracker data. Our solution measured the impact of: 

  • Online advertising (YouTube, Facebook, and Twitter ads; AdWords)
  • Offline advertising (TV ads; in-store coupons)
  • Promotional Pricing and Promotions
  • Positive and negative halo effect across product categories
  • Baseline components: base price, assortment, market penetration, seasonality, holidays, geographical location, and brand awareness


Result: Calculating ROMI Drives Marketing Activities and Yield Higher Revenue

As a result of the modelling, we were able to create a budget optimizer tool that measured market performance and calculated the optimal ROMI, resulting in a budget allocation decision that was uploaded into the tool that the client used to allocate marketing spend by geography, product, and channel. 

Our work allowed the client to fully understand the return on their marketing investment for all the channels they were spending money on resulting in a $1.2MM advertising savings and decreased frequency of discounting. We also decreased cross-category cannibalization from 86% to 37%. The impact on sales was evident with 22% increase in margin.