Predicting In-Store Promotion Effectiveness Linked To Macro Marketing Campaigns

Predicting In-Store Promotion Effectiveness Linked To
Macro Marketing Campaigns

The Challenge

A leading electronics retailer invests in marketing to increase the footfall in their stores. In-store promotions then try to convert visitors to buyers and increase the value of their basket. The company wanted to predict the effectiveness of different campaigns given the marketing campaigns they have run in the recent past. They tasked Tatras to build the model so that they could optimize promotions in-store to maximize the returns from marketing campaigns that drove the footfall.

Hypothesis

  • Historical sales and marketing campaign data can be used to predict the demographic of the footfall in store, given a set of campaigns.
  • Campaigns run in the recent past affect the demographic of the footfall and hence should drive in-store promotions.
  • In-store promotions that are personalized to the demographic of the footfall will lead to more sales revenue.

Execution

  • Forecast Sales for individual products and product groups/categories, given marketing campaign history.
  • Incorporate promotion attributes into each of the forecasts.
  • Measure halo, pull forward, cannibalization to compute item’s market share and revenue lift.
  • Measure statistical significance of effect of promotion on market share and revenue.

Outcomes

  • The model was deployed in production and is used to design promotions instore.

Project Highlights

82%

ACCURACY IN SALES FORECASTS

13%

LIFT IN SALES ATTRIBUTED TO INSTORE PROMOTIONS