Joint Product Framing and Order Fulfillment for E-Commerce Retailers
Co-authored with Yanzhe Lei (Queen's University, Canada), Joline Uichanco (Univ of Michigan, US), Andrew Vakhutinsky (Oracle Lab, US)
How should an online retailer decide the set of products to be displayed for promotion, the set of products to be placed at the top of a display page, etc.? These are examples of so-called "product framing". It has been widely noted in the empirical literature that product framing (i.e., display, ranking, pricing) matters. They affect customers' attention, which in turn affect their purchasing decision. This talk presents a recent joint work on the topic of randomized product framing and order fulfillment for e-commerce retailers. We analyze a relatively general setting in which customers arrive sequentially over time and the retailer's objective is to maximize his expected total profits throughout a finite selling horizon (e.g., clearance sales periods). The technical challenge of the problem comes from the fact that, in practice, the retailer's decision must satisfy a certain set of constraints including the inventory constraints (i.e., cumulative sales of a product cannot exceed its available inventory, or else there will a penalty) and the cardinality constraints (e.g., at most 20 products can be displayed for promotion at any given time). We develop a heuristic policy that can be theoretically shown to be asymptotically optimal in the setting with a large demand and large inventory. We also numerically test our heuristic policy using both synthetic and real-world data provided by a major US retailer. The results show that the proposed heuristic is very close to optimal and also outperforms many benchmarks and some state-of-the-art algorithms.