Dynamic Pricing of Omnichannel Inventories

Abstract

Omnichannel retail refers to a seamless integration of the e-commerce channel and the network of brickand-mortar stores. New cross-channel interactions emerge from the integration: (i) cross-channel fulfillment allows inventory to be shared between channels, and (ii) accessible price information induces price-based channel substitution. However, existing price optimization systems have not been able to keep pace with these new interactions. In this paper, we propose a novel multistage stochastic program for the dynamic pricing of a product offered in an omnichannel network incorporating the new interactions. We propose a deterministic and a robust heuristic where, in each period, omnichannel prices and cross-channel fulfillment inventories are jointly optimized via computationally tractable mixed integer programs. In extensive simulations, the heuristics incur an average revenue loss of less than 4% and 8%, respectively. The key benefits of these approaches arise from inventory rebalancing using omnichannel prices (more cross-channel fulfillment inventories from stores with low sell-through rates) and through better management of channel demands. In experiments with historical consumer electronics data of a major U.S. retailer, the proposed analytics shows a 6-12% increase in markdown revenue over the retailer's actual revenue across different categories. A proprietary implementation of the analytics is now commercially available as part of the IBM Commerce Markdown Price Solution and has been piloted for clearance pricing at the retailer. A causal model analysis on the live pilot data shows a 12% increase in clearance period revenue.

Why Empirical Research Is Good for Operations Management, and What Is Good Empirical Operations Management?

“…Musalem et al (2010) and Vulcano et al (2012) develop methods to estimate these types of models when the demand data are censored because of product stock-outs. There is now extensive research developing decision models for assortment, inventory, and pricing optimization using different variations of RUMs (Davis et al 2014, Rusmevichientong andShmoys 2014), some of which have been applied in practice (Fisher and Vaidyanathan 2014;Harsha et al 2019).…”

Data Analytics in Operations Management: A Review

“…In the omnichannel retail setting, consumers can compare store prices with online prices, giving rise to challenging pricing problems. Harsha et al (2019) study the price optimization aspect of omnichannel operations. Due to the presence of cross-channel interactions in demand and supply, the paper proposes two pricing policies that are based on the idea of "partitions" to the store inventory, which approximate how it will be used by the two different channels.…”

Frontiers in Service Science: The Management of Data Analytics Services: New Challenges and Future Directions

“…Using data from a large online retailer, they propose a method that reduces outbound shipping costs on the order of 1%. Harsha et al (2019) enhance this problem by considering the omnichannel price optimization problem, that is, pricing in the presence of cross-channel interactions in demand and supply, with exogenous cross-channel fulfillment. They implemented their approach commercially as part of the IBM Commerce Markdown Price Solution.…”