Our client, with presence in close to 6000+ locations, wanted to cross-sell their chicken wings offerings to their pizza customers.
They wanted to understand the best customers to reach out to based on their propensity to buy.
With a low marketing budget, the client wanted to use digital marketing to reach out to these customers.
We followed a step-wise statistical approach on client-provided data to segment its stores and identify opportunity areas to increase sales.
Step 1: Developed Propensity Modeling
Multinomial logistic regression was the method used to develop a model with a significantly improved lift curve. This helped to score customers on propensity.
Step 2: Understand regional priorities and digital channel propensity
- By combining mosaic data, we identified Key Designated Market Areas (DMAs) with high potential.
- Channel response propensity was developed based on the demographic characteristics.
Step 3: Define Targeted Marketing Strategy for each segment
We combined the results of channel propensity with the purchase propensity to define the marketing strategy.
We could identify DMAs with high potential.
A Channel Response propensity framework helped in identifying suitable channels for targeting customers.
We identified propensity to buy for each of the approximately 100m customers.
Suitability of digital messaging for each customer was identified and suitable offer categories for each consumer type were recommended.