Media Mix Optimization
Create the optimum media channel spend to maximize RoI.
Background & Challenges
One of the leading North American Retailers in US wanted to
- Quantify Media Drivers of Short Term Sales Performance
- Compute Marginal return on investment figures for each Media Channel
- Optimise Spend on each Media Driver
PHASE I: MODELING
Media Channels are modeled as Fixed Effect, Random Effect & Fixed Random Effects
PHASE II: AROI vs. MROI
We compare the RoI based on diminishing law of returns. As an implication, we need to look for the threshold beyond which, even though ARoI increases, MRoI begins to provide lower returns.
PHASE III: MEDIA CHANNEL OPTIMIZATION
- Unobserved Component Modelling: Decompose Total Sales to generate time varying stochastic base
- Panel Data Modelling : for Incremental Sales for each Media Channels
Salest = β0Baset+ β1 x1t+ β2 x2t+… + εt
- Transformations: The Marketing Channels are transformed to reflect their behaviour in marketing. This involves creating Ad stock, Power & Lag transformations.
- Treated as a non-linear optimization problem
- Provide reallocation spend recommendations, maximizing the returns
- Constraints: individual & overall spend constraints
Optimum reallocation spend recommendations of each media driver, maximising the returns for the budget quarter.