Optimize Promotions

Promotion Analysis: It is critical that the retailer understand the characteristics of its baskets containing promoted items. If unusually high-profit baskets tend to contain certain items when they are promoted, one can infer that the promotion increases short-term profitability. If predominately low-profit, cherry picker baskets contain a promoted item, one can infer that the promotion fails to have an immediate positive effect on profit.

We can apply market basket analysis to study the longer-term effects of a promotion: does number of baskets increase during the promotion (traffic driver), does the increase in number of baskets attenuate as the promotion grows old or after it is discontinued, how does the promotion’s effect on basket composition change over time? Note that it is important that this analysis consider promotion type (e.g., flyer, in-store display) as well as depth of price reduction. Learnings resulting from this analysis can be made actionable via incorporation into PriceStrat’s demand models and promotional components, or into any other process the retailer might employ for these purposes.

Tailored Offers: Based on the segmentation and analysis of shopping behavior, we can design sets of offers that are collectively more effective than each individual offer taken alone. This could involve simultaneous offers on sets of similar items for some consumers, and offers that span many purchase categories for other consumers. We have developed a supervised learning-based system which forecasts from consumer past shopping behavior how receptive they are likely to be to redeem offers for new items.