An International Beverage & Brewing Company Appreciated An Estimated $300 Mn Annual Impact By Optimizing Promotional Mix & Spending

A leading beverage and brewing company, catering to customers across the globe, wanted to understand the efficacy of its promotions. The company primarily used intuition in planning the promotional mix and wanted to move to insight-driven decision-making in optimizing the promotional mix, economizing spending and maximizing returns.

TheMathCompany interpreted and bucketed historical promotions attributes, measured multiple KPIs, and developed insights on promotion efficacy to optimize spend on promotions.

The client wished to identify a scientific method to measure the success/failure of promotions and thereby, find the most effective, optimal mix of promotions. We first looked through, interpreted and bucketed the traits of historical promotions. Then KPIs were created to devise the effectiveness and impact of different promotions, and multi-level insights were generated. Step-wise Solutioning to Optimize Promotions Mix & Spending. Attributes such as price effect, quantity effect, merchandise effect, mixed effect, channel mix, placement mix, messaging, time period, discount bucket etc., were used to isolate overlap of type and timeline of promotions. Contribution methodologies (UCM, Mix, regression etc.) were codified to determine attribute impact. MMX models were built to determine short-term and long-term marketing impact, and brand impact isolation using hierarchical multi-stage UCM model. Latency was reduced by tapping into insights on campaigns in real-time as opposed to quarterly or half-yearly insights. Promotions were measured on multiple KPIs such as volume impact, marginal contribution, new launch lift, shopper reach, share efficiency, seasonal lift etc. to allow prioritization and alignment with product strategy. Econometric modelling was used for portfolio optimization through a holistic view of risk vs. return of investing in multiple countries (MCMC techniques). Need-based customer analysis was conducted using state-space modelling to predict propensity and drivers for progressing in the states value chain.