Yossi Aviv

The Effect of Collaborative Forecasting on Supply Chain Performance

We consider a cooperative, two-stage supply chain consisting of two members: a retailer and a supplier. In our first model, called local forecasting, each member updates the forecasts of future demands periodically, and is able to integrate the adjusted forecasts into his replenishment process. Forecast adjustments made at both levels of the supply chain can be correlated. The supply chain has a decentralized information structure, so that day-to-day inventory and forecast information are known locally only. In our second model, named collaborative forecasting, the supply chain members jointly maintain and update a single forecasting process in the system. Hence, forecasting information becomes centralized. Finally, we consider as a benchmark the special case in which forecasts are not integrated into the replenishment processes at all. We propose a unified framework that allows us to study and compare the three types of settings. This study comes at a time when various types of collaborative forecasting partnerships are being experimented within industry, and when the drivers for success or failure of such initiatives are not yet fully understood. In addition to providing some managerial insights into questions that arise in this context, our set of models is tailored to serve as building blocks for future work in this emerging area of research.

Optimal Pricing of Seasonal Products in the Presence of Forward-Looking Consumers

We study the optimal pricing of a finite quantity of a fashion-like seasonal good in the presence of forward-looking (strategic) customers. We distinguish between two classes of pricing strategies: contingent and announced fixed-discount. In both cases, the seller acts as a Stackelberg leader announcing his pricing strategy, while consumers act as followers taking the sellers strategy as given and determining their purchasing behavior. In each case, we identify a subgame-perfect Nash equilibrium and show that given the sellers strategy, the equilibrium in the consumer subgame is unique and consists of symmetric threshold purchasing policies. For both cases, we develop a benchmark model in which customers are nonstrategic (myopic). We conduct a comprehensive numerical study to explore the impact of strategic consumer behavior on pricing policies and expected revenue performance. We show that strategic customer behavior suppresses the benefits of price segmentation, particularly under medium-to-high values of heterogeneity and modest rates of decline in valuations. However, when the level of consumer heterogeneity is small, the rate of decline is medium-to-high, and the seller can optimally choose the time of discount in advance, segmentation can be used quite effectively even with strategic consumers. We find that the seller cannot avoid the adverse impact of strategic consumer behavior even under low levels of initial inventory. We argue that while the seller expects customers to be more concerned about product availability at discount time, he cannot use high-price “betting” strategies as he would in the case of low inventory and myopic customers. Under certain qualifications, announced fixed-discount strategies perform essentially the same as contingent pricing policies in the case of myopic consumers. However, under strategic consumer behavior, announced pricing policies can be advantageous to the seller, compared to contingent pricing schemes. Interestingly, those cases that announced discount strategies offer a significant advantage compared to contingent pricing policies. They appear to offer only a minimal advantage in comparison to fixed-pricing policies. Finally, when the seller incorrectly assumes that strategic customers are myopic in their purchasing decisions, it can be quite costly, reaching potential revenue losses of about 20%.

Gaining Benefits from Joint Forecasting and Replenishment Processes: The Case of Auto-Correlated Demand

In this paper we consider a cooperative, two-level supply chain consisting of a retailer and a supplier. As in many practical settings, the supply chain members progressively observe market signals that enable them to explain future demand. The demand itself evolves according to an auto-regressive time series. We examine three types of supply chain configurations. In the first setting, the retailer and the supplier coordinate their policy parameters in an attempt to minimize systemwide costs, but they do not share their observations of market signals. In the second setting, resembling many original vendor-managed inventory (VMI) programs, the supplier takes the full responsibility of managing the supply chains inventory, but the retailers observations of market signals are not transferred to him. In our third setting, reminiscent of collaborative forecasting and replenishment partnerships, inventory is managed centrally, and all demand related information is shared. We propose a set of stylized models to study the three settings and use them to provide managerial insights into the value of information sharing, VMI, and collaborative forecasting.

On the Benefits of Collaborative Forecasting Partnerships Between Retailers and Manufacturers

This paper studies the potential benefits of collaborative forecasting (CF) partnerships in a supply chain that consists of a manufacturer and a retailer. To reflect the reality in production environments, we propose a scorecard that captures inventory considerations, production smoothing, and adherence-to-plans. We present a prescriptive convex-cost production planning model for the manufacturer, and a replenishment model for the retailer. We use our integrative reference model to study the potential benefits of CF partnerships. Overall, we find that the benefits of CF depend on the following key characteristics of the supply chain: the relative explanatory power of the supply chain partners, the supply side agility, and the internal service rate. CF is expected to bring high benefits to the supply chain when the manufacturer has the largest relative explanatory power. But quite disappointingly, in these cases a CF partnership does not appear to be valuable to the manufacturer. When the retailer is the dominant observer of market signals, CF typically yields a “win-win” outcome. In order to effectively act upon the information exchanged via CF, the supply side needs to be sufficiently agile. The benefits reported in this paper should be considered as conservative. This is because CF partnerships often bring better information, improved decision support technologies, as well as process improvement to the trading partners. Consequently, the supply side agility can be improved. If this indeed happens, the compound benefits of CF can be dramatically higher than our conservative estimates. Finally, we provide a qualitative discussion of the possible role of internal service rates in supply chains, either as planning parameters to improve performance, or as a mechanism for sharing the benefits of CF between the trading partners.

A Time-Series Framework for Supply-Chain Inventory Management

We consider a supply chain in which the underlying demand process can be described in a linear state space form. Inventory is managed at various points of the chain (members), based on local information that each member observes and continuously updates. The key feature of our model is that it takes into account the ability of the members to observe subsets of the underlying state vector, and adopt their forecasting and replenishment policies accordingly. This enables us to model situations in which the members are privy to certain explanatory variables of the demand, with the latter possibly evolving according to a vector autoregressive time series. For each member, we identify an associated demand evolution model, for which we propose an adaptive inventory replenishment policy that utilizes the Kalman filter technique. We then provide a simple methodology for assessing the benefits of various types of information-sharing agreements between members of the supply chain. We also discuss inventory positioning and cost performance assessment in the supply chain. Our performance metrics and inventory target levels are usually presented in matrix forms, allowing them to accommodate a relatively large spectrum of linear demand models, and making them simple to implement. Several illustrations for possible applications of our models are provided.

Design for Postponement: A Comprehensive Characterization of Its Benefits Under Unknown Demand Distributions

Recent papers have developed analytical models to explain and quantify the benefits of delayed differentiation and quick response programs. These models assume that while demands in each period are random, they are independent across time and their distribution is perfectly known, i.e., sales forecasts do not need to be updated as time progresses. In this paper, we characterize these benefits in more general settings, where parameters of the demand distributions fail to be known with accuracy or where consecutive demands are correlated. Here it is necessary to revise estimates of the parameters of the demand distributions on the basis of observed demand data. We analyze these systems in a Bayesian framework, assuming that our initial information about the parameters of the demand distributions is characterized via prior distributions. We also characterize the structure of close-to-optimal ordering rules in these systems, for a variety of types of order cost functions.

Optimal Markdown Pricing: Implications of Inventory Display Formats in the Presence of Strategic Customers

We propose a game-theoretical model of a retailer who sells a limited inventory of a product over a finite selling season by using one of two inventory display formats: display all (DA) and display one (DO). Under DA, the retailer displays all available units so that each arriving customer has perfect information about the actual inventory level. Under DO, the retailer displays only one unit at a time so that each customer knows about product availability but not the actual inventory level. Recent research suggests that when faced with strategic consumers, the retailer could increase expected profits by making an upfront commitment to a price path. We focus on such pricing strategies in this paper, and study the potential benefit of DO compared to DA, and its effectiveness in mitigating the adverse impact of strategic consumer behavior. We find support for our hypothesis that the DO format could potentially create an increased sense of shortage risk, and hence it is better than the DA format. However, although potentially beneficial, a move from DA to DO is typically very far from eliminating the adverse impact of strategic consumer behavior. We observe that, generally, it is not important for a retailer to modify the level of inventory when moving from a DA to a DO format; a change in the display format, along with an appropriate price modification, is typically sufficient. Interestingly, across all scenarios in which a change in inventory is significantly beneficial, we observed that only one of the following two actions takes place: either the premium price is increased along with a reduction in inventory, or inventory is increased along with premium price reduction. We find that the marginal benefit of DO can vary dramatically as a function of the per-unit cost to the retailer. In particular, when the retailers per-unit cost is relatively high, but not too high to make sales unprofitable or to justify exclusive sales to high-valuation customers only, the benefits of DO appear to be at their highest level, and could reach up to 20% increase in profit. Finally, we demonstrate that by moving from DA to DO, while keeping the price path unchanged, the volatility of the retailers profit decreases.

A Partially Observed Markov Decision Process for Dynamic Pricing

In this paper, we develop a stylized partially observed Markov decision process (POMDP) framework to study a dynamic pricing problem faced by sellers of fashion-like goods. We consider a retailer that plans to sell a given stock of items during a finite sales season. The objective of the retailer is to dynamically price the product in a way that maximizes expected revenues. Our model brings together various types of uncertainties about the demand, some of which are resolvable through sales observations. We develop a rigorous upper bound for the sellers optimal dynamic decision problem and use it to propose an active-learning heuristic pricing policy. We conduct a numerical study to test the performance of four different heuristic dynamic pricing policies in order to gain insight into several important managerial questions that arise in the context of revenue management.

Capacitated Multi-Item Inventory Systems with Random and Seasonally Fluctuating Demands: Implications for Postponement Strategies

We address multi-item inventory systems with random and seasonally fluctuating, and possibly correlated, demands. The items are produced in two stages, each with its own lead-time; in the first stage a common intermediate product is manufactured. The production volumes in the first stage are bounded by given capacity limits. We develop an accurate lower bound and close-to-optimal heuristic strategies of simple structure. The gap between them, evaluated in an extensive numerical study, is on average only 0.45%. We use the model to investigate the benefits of various delayed product differentiation (postponement) strategies, as well as other strategic questions, including (i) the benefits of flexible versus dedicated production facilities; (ii) the trade-off between capacity and inventory investments; and (iii) the trade-off between capacity investments and service levels.

Counteracting Strategic Consumer Behavior in Dynamic Pricing Systems

Dynamic pricing and revenue management practices are gaining increasing popularity in the retail industry, and have engendered a large body of academic research in recent decades. When applying dynamic pricing systems, retailers must account for the fact that, often, strategic customers may time their purchases in anticipation of future discounts. Such strategic consumer behavior might lead to severe consequences on the retailers’ revenues and profitability. Researchers have explored several approaches for mitigating the adverse impact of this phenomenon, such as rationing capacity, making price and capacity commitments, using internal price-matching policies, and limiting inventory information. In this chapter, we present and discuss some relevant theoretical contributions in the management science literature that help us understand the potential value of the above mitigating strategies.