Engineering Social Learning: Information Design of Time-Locked Sales Campaigns for Online Platforms

Many online platforms offer time-locked sales campaigns, whereby products are sold at fixed prices for prespecified lengths of time. Platforms often display some information about previous customers’ purchase decisions during campaigns. Using a dynamic Bayesian persuasion framework, we study how a revenue-maximizing platform should optimize its information policy for such a setting. We reformulate the platform’s problem equivalently by reducing the dimensionality of its message space and proprietary history. Specifically, three messages suffice: a neutral recommendation that induces a customer to make her purchase decision according to her private signal about the product and a positive (respectively (resp.), negative) recommendation that induces her to purchase (resp., not purchase) by ignoring her signal. The platform’s proprietary history can be represented by the net purchase position, a single-dimensional summary statistic that computes the cumulative difference between purchases and nonpurchases made by customers having received the neutral recommendation. Subsequently, we establish structural properties of the optimal policy and uncover the platform’s fundamental trade-off: long-term information (and revenue) generation versus short-term revenue extraction. Further, we propose and optimize over a class of heuristic policies. The optimal heuristic policy provides only neutral recommendations up to a cutoff customer and provides only positive or negative recommendations afterward, with the recommendation being positive if and only if the net purchase position after the cutoff customer exceeds a threshold. This policy is easy to implement and numerically shown to perform well. Finally, we demonstrate the generality of our methodology and the robustness of our findings by relaxing some informational assumptions. This paper was accepted by Gabriel Weintraub, revenue management and market analytics.

A Simple Heuristic Policy for Stochastic Distribution Inventory Systems with Fixed Shipment Costs

Inventory Management in Stochastic Distribution Systems

When Variability Trumps Volatility: Optimal Control and Value of Reverse Logistics in Supply Chains with Multiple Flows of Product

Problem definition: We study how to optimally control a multistage supply chain in which each location can initiate multiple flows of product, including the reverse flow of orders. We also quantify the resulting value generated by reverse logistics and identify the drivers of that value. Academic/practical relevance: Reverse logistics has been gaining recognition in practice and theory for helping companies better match supply with demand, and thus reduce costs in their supply chains. Nevertheless, there remains a lack of clarity in practice and the research literature regarding precisely what in reverse logistics is so important, exactly how reverse logistics creates value, and what the drivers of that value are. Methodology: We first formulate a multistage inventory model to jointly optimize ordering decisions pertaining to regular, reverse, and expedited flows of product in a logistics supply chain, where the physical transformation of the product is completed at the most upstream location. With multiple product flows, the feasible region for the problem acquires multidimensional boundaries that lead to the curse of dimensionality. Next, we extend our analysis to product-transforming supply chains, in which product transformation is allowed to occur at each location. In such a system, it becomes necessary to keep track of both the location and stage of completion of each unit of inventory; thus, the number of state and decision variables increases with the square of the number of locations. Results: To solve the reverse logistics problem in logistics supply chains, we develop a different solution method that allows us to reduce the dimensionality of the feasible region and identify the structure of the optimal policy. We refer to this policy as a nested echelon base stock policy, as decisions for different product flows are sequentially nested within each other. We show that this policy renders the model analytically and numerically tractable. Our results provide actionable policies for firms to jointly manage the three different product flows in their supply chains and allow us to arrive at insights regarding the main drivers of the value of reverse logistics. One of our key findings is that, when it comes to the value generated by reverse logistics, demand variability (i.e., demand uncertainty across periods) matters more than demand volatility (i.e., demand uncertainty within each period). To analyze product-transforming supply chains, we first identify a policy that provides a lower bound on the total cost. Then, we establish a special decomposition of the objective cost function that allows us to propose a novel heuristic policy. We find that the performance gap of our heuristic policy relative to the lower-bounding policy averages less than 5% across a range of parameters and supply chain lengths. Managerial implications: Researchers can build on our methodology to study more complex reverse logistics settings, as well as tackle other inventory problems with multidimensional boundaries of the feasible region. Our insights can help companies involved in reverse logistics to better manage their orders for products, and better understand the value created by this capability and when (not) to invest in reverse logistics.

Pricing and Matching with Forward-Looking Buyers and Sellers

Problem definition: We study a dynamic market over a finite horizon for a single product or service in which buyers with private valuations and sellers with private supply costs arrive following Poisson processes. A single market-making intermediary decides dynamically on the ask and bid prices that will be posted to buyers and sellers, respectively, and on the matching decisions after buyers and sellers agree to buy and sell. Buyers and sellers can wait strategically for better prices after they arrive. Academic/practical relevance: This problem is motivated by the emerging sharing economy and directly speaks to the core of operations management that is about matching supply with demand. Methodology: The dynamic, stochastic, and game-theoretic nature makes the problem intractable. We employ the mechanism-design methodology to establish a tractable upper bound on the optimal profit, which motivates a simple heuristic policy. Results: Our heuristic policy is: fixed ask and bid prices plus price adjustments as compensation for waiting costs, in conjunction with the greedy matching policy on a first-come-first-served basis. These fixed base prices balance demand and supply in expectation and can be computed efficiently. The waiting-compensated price processes are time-dependent and tend to have opposite trends at the beginning and end of the horizon. Under this heuristic policy, forward-looking buyers and sellers behave myopically. This policy is shown to be asymptotically optimal. Managerial implications: Our results suggest that the intermediary might not lose much optimality by maintaining stable prices unless the underlying market conditions have significantly changed, not to mention that frequent surge pricing may antagonize riders and induce riders and drivers to behave strategically in ways that are hard to account for with traditional pricing models.

Emergency Response Optimization using Online Hybrid Planning

This paper poses the planning problem faced by the dispatcher responding to urban emergencies as a Hybrid (Discrete and Continuous) State and Action Markov Decision Process (HSA-MDP). We evaluate the performance of three online planning algorithms based on hindsight optimization for HSA- MDPs on real-world emergency data in the city of Corvallis, USA. The approach takes into account and respects the policy constraints imposed by the emergency department. We show that our algorithms outperform a heuristic policy commonly used by dispatchers by significantly reducing the average response time as well as lowering the fraction of unanswered calls. Our results give new insights into the problem such as withholding of resources for future emergencies in some situations.