Using Big Data to Predict Consumer Responses to Promotional Discounts as Part of Sales & Operations Planning

Abstract

Price promotions (discounts) are a well-known means by which a supply chain can stimulate demand for a product. These promotions could affect demand for a product in three ways by: 1) increasing the overall market growth, 2) stealing market share from competitors, and/or 3) increasing the amount of consumer forward buying. Supply chain members must be able to estimate these effects on demand and the corresponding effects on both revenues and costs when conducting sales and operations planning. We analyzed the effects on demand using a big data approach on promotional data made publicly available by Grupo Bimbo (a multinational bakery product manufacturing company headquartered in Mexico City). This company offered promotional coupons to customers for particular items. Bimbo captured sales history for each customer on how often they shopped, what they bought, and the amount that they spent. Bimbo then tracked how many times during the next year that customers returned to buy the promoted items at full price. Using this data set, we assessed which types of offers were more effective at achieving the goal of increasing repeat purchases at full price. Whether the offer was for a weekend or weekday had no significant effect. However, we found that a larger discount percent was associated with fewer repeat purchases at full price. Further, customers who tended to spend more, on average, per trip had a higher number of repeat purchases for an item.

Keywords: Big Data, Data Analytics, Multiple Regression, Promotions, Sales and Operations Planning

JEL Classification: C33, C35, M110