The U.S. economy is already dominated by service and information-intensive industries in terms of both gross national product and jobs, and these trends are visible in all major world economies. These economic shifts are driven by productivity changes, which today often depend on new information and communication technologies. These changes can be thought of as service industrialization, which underlies productivity improvements. Industrialization is, in turn, closely related to the design and operation of service processes at the level of firms and sectors. Some of the implications for process economics, operations strategy, and process management are outlined, and the opportunities for research in operations and technology management related to these trends are discussed.

This paper compares the commonly used periodic review, replenishment interval, order-up-to (R, T ) policy to the continuous review, reorder point, order quantity, (Q, r) model. We show that long-run average cost function for the single-product (R, T ) policy has a structure similar to that of the (Q, r) model. Consequently, many of the useful properties of the latter model are applicable. In particular, the optimal cost is insensitive to the choice of the replenishment interval, T, provided the optimal order-up-to level, R, corresponding to T is used. For instance, a suboptimal T obtained from a deterministic analysis increases costs by no more than 6.125%. For continuous demand, we analytically prove that use of a (R, T)policy instead of the optimal policy increases costs by at most 41.42% in the worst case. Computational experiments on Poisson demand demonstrate that the average-case relative error of using a (R,T)policy is under 7.5%. This relative error is lower when the demand rate and leadtime are high and the fixed order costs are either very low or very high. When coordination of order placement epochs is desirable, the (R,T) policy may sometimes be preferred to the (Q, r) policy. In this context, we illustrate application of our single-product results to more complex systems. In particular, we show that a simple power-of-two, (R,T) based heuristic for the stochastic multiproduct joint replenishment problem has a worst-case performance guarantee of 1.5. A similar result is explored for a special case of a two-echelon serial inventory system

In a firm that makes schedule, orders are always processed within a fixed time frame. In a congested facility, such a situation would be impossible using conventional queuing logic. We propose a conceptual model of a firm in which workers make schedule by rushing jobs, if necessary, with potential quality consequences. Hence, time and quality are substitutes, a feature that we recognize explicitly in our definition of the firm's capacity. Production yields are not exogenous parameters but endogenously determined by workers responding to schedule pressures. The plant manager can authorize overtime to relieve this pressure or live with the quality consequences of rushing. The model reveals the close relationships among the firm's workforce policies, the integrity of the inspection system, and the cost performance of the firmas its volume and/or product line expands. We consider pure congestion (driven by volume) and pure complexity (driven by product line breadth) effects in the context of our plant performance model. We consider two root causes of complexity costs: time and quality. The time effects of complexity will only have cost consequences in congested facilities, but quality effects are always present. Also in contrast to time effects, quality effects of complexity can be present in nonbottleneck workstations. Hence, the quality consequences of complexity can be as, or more important than the time consequences.

Many firms face environments with long manufacturing leadtimes, great product variety, and uncertain, nonstationary demand. A challenge is how to plan production and inventories to provide the best customer service at the least cost. In this paper, we first describe an application at Teradyne in which we implemented an inventory hedge to protect against cyclic demand variability. Based on this experience, we develop a model to better understand the efficacy of this hedging policy. We consider an inventory system for a single aggregate product with a Markov-modulated Poisson demand process. We provide approximate performance measures for this system and develop an optimization problem for determining the size and location of an intermediate-decoupling inventory. We use this optimization to show the value of an intermediate-decoupling inventory as a hedge for cyclic demand environments.

The revolution in information technology has provided the research community in operations management (OM) with new areas to explore and many new avenues to develop. In recognition of this, many editors of OM journals strive to publish new ideas. These two forces generate strong motivation, but many young OM scholars find it difficult to find new OM research ideas. To address this challenge, I describe three simple thought processes (or approaches) that I have learned and used to identify new research topics over the last 35 years. These approaches are as follows: (1) observe and learn to develop “problem-based” research, (2) ask “whys” to develop “phenomenon-based” research, and (3) sketch graphs to develop “insight-based” research. Clearly, these simple approaches are neither complete nor optimal; however, I share these personal thought processes with the hope of contributing to discussions as to how the next generation of OM researchers can build upon and expand the remarkable impact that our field has had and will continue to have.

Should OM research become more relevant? I provide my own perspective on this question and some practical ideas for our community to explore. This article is based on my Manufacturing and Service Operations Management (MSOM) Distinguished Fellow inaugural lecture given at the University of Toronto on June 29, 2015.

Problem definition: Traditional monopoly pricing assumes sellers have full information about consumer valuations. We consider monopoly pricing under limited information when a seller only knows the mean, variance, and support of the valuation distribution. The objective is to maximize expected revenue by selecting the optimal fixed price. Methodology/results: We adopt a distributionally robust framework, in which the seller considers all valuation distributions that comply with the limited information. We formulate a maximin problem that seeks to maximize expected revenue for the worst case valuation distribution. The minimization problem that identifies the worst case valuation distribution is solved using primal-dual methods and, in turn, leads to an explicitly solvable maximization problem. This yields a closed-form optimal pricing policy and a new fundamental principle prescribing when to use low and high robust prices. Managerial implications: We show that the optimal policy switches from low to high prices when variance becomes sufficiently large, yielding significant performance gains compared with existing robust prices that generally decay with market uncertainty. This presents guidelines for when the seller should switch from targeting mass markets to niche markets. Similar guidelines are obtained for delay-prone services with rational utility-maximizing customers, underlining the universality and wide applicability of the novel pricing policy.

Two innovative financing schemes have emerged in recent years to enable suppliers to obtain financing for production. The first, purchase order financing (POF), allows financial institutions to offer loans to suppliers by considering the value of purchase orders issued by reputable buyers. Under the second, which we call buyer direct financing (BDF), manufacturers issue both sourcing contracts and loans directly to suppliers. Both schemes are closely related to the supplier’s performance risk (whether the supplier can deliver the order successfully), upon which the repayment of these loans hinges. To understand the relative efficiency of the two emerging schemes, we analyze a game-theoretical model that captures the interactions among three parties (a manufacturer, a financially constrained supplier who can exert unobservable effort to improve delivery reliability, and a bank). We find that, when the manufacturer and the bank have symmetric information, POF and BDF yield the same payoffs for all parties irrespective of the manufacturer’s control advantage under BDF. The manufacturer, however, has more flexibility under BDF in selecting contract terms. In addition, even when the manufacturer has superior information about the supplier’s operational capability, the manufacturer can efficiently signal her private information via the sourcing contract if the supplier’s asset level is not too low. As such, POF remains an attractive financing option. However, if the supplier is severely financially constrained, the manufacturer’s information advantage makes BDF the preferred financing scheme when contracting with an efficient supplier. In particular, the relative benefit of BDF (over POF) is more pronounced when the supply market contains a larger proportion of inefficient suppliers, when differences in efficiency between suppliers are greater, or when the manufacturer’s alternative sourcing option is more expensive. The online appendix is available at https://doi.org/10.1287/msom.2017.0638 . This paper has been accepted for the Manufacturing & Service Operations Management Special Issue on Interface of Finance, Operations, and Risk Management.

In the early 1990s, research began to show that the Japanese production theory, which espouses reduction of machine setup time as a sure way to improve production performance, may be limited. Specifically, it was found that reduction in mean setup times without any change in variance can, paradoxically, increase waiting time and work in process (WIP) in a cyclic production system. Setup time variance was demonstrated to play a central role because of the paradoxes it produced with the resulting harm to effective capacity. Subsequently, explicit formulas were derived for determining whether adding fixed forced idle time (but holding variance constant) would reduce waiting time and, if so, the optimal amount of idle time to add. However, research to date has offered little guidance to reduce setup time variance to improve waiting time. We show that a greater reduction is achievable by adding a variable idle time that is a nonincreasing function of setup time and thereby reduce the combined setup time variance. We provide explicit procedures for finding the optimal variable idle time as a function of setup time when the latter follows any finite discrete distribution. We also show how to implement our policy and show that our approach can improve waiting time even when other currently known approaches cannot.