Knowledge that Transforms

In a call center, staffing decisions must be made before the call arrival rate is known with certainty. Once the arrival rate becomes known, the call center may be over‐staffed, in which case staff are being paid to be idle, or under‐staffed, in which case many callers hang‐up in the face of long wait times. Firms that have chosen to keep their call center operations in‐house can mitigate this problem by co‐sourcing; that is, by sometimes outsourcing calls. Then, the required staffing N depends on how the firm chooses which calls to outsource in real time, after the arrival rate realizes and the call center operates as a M / M / N + M queue with an outsourcing option. Our objective is to find a joint policy for staffing and call outsourcing that minimizes the long‐run average cost of this two‐stage stochastic program when there is a linear staffing cost per unit time and linear costs associated with abandonments and outsourcing. We propose a policy that uses a square‐root safety staffing rule, and outsources calls in accordance with a threshold rule that characterizes when the system is “too crowded.” Analytically, we establish that our proposed policy is asymptotically optimal, as the mean arrival rate becomes large, when the level of uncertainty in the arrival rate is of the same order as the inherent system fluctuations in the number of waiting customers for a known arrival rate. Through an extensive numerical study, we establish that our policy is extremely robust. In particular, our policy performs remarkably well over a wide range of parameters, and far beyond where it is proved to be asymptotically optimal.

As the Obama administration steps up oversight of high-risk IT projects, contracting organizations must take greater responsibility to provide a level of confidence in the services they offer. That is where one of the latest offerings from the Software Engineering Institute at Carnegie Mellon University can help. (Sacks 2010). A number of US federal government information technology (IT) initiatives (e.g., implementation of the Health Insurance Marketplace or Obamacare, development of navigation systems in missiles and unmanned vehicles) are frequently organized in the form of IT projects (Kundra 2010). The Office of Management and Budget (OMB), which tracks the progress of all federal IT projects indicates that such projects face significant schedule and cost overruns. Nearly 25% of federal IT projects with a cumulative budget exceeding $10 billion and spread over 28 government agencies are facing moderate to severe problems in meeting their schedule and budgetary targets (Source: www.itdashboard.gov). Additionally, the US Government Accountability Office (GAO) reports that nearly 72% of federal IT projects, with a total budget of $27 billion, are poorly planned and face significant schedule and cost overruns (US GAO Report 2010). Despite the evidence of schedule and budget overruns in federal IT projects, the challenges associated with the management of such projects, and more generally, of IT projects in the public sector, have received limited attention in both practice and research. Public sector projects differ from private sector projects in a number of ways (Boyne 2002). First, federal IT projects are primarily funded with taxpayer's money and are aimed at maximizing public utility, instead of maximizing profits, as in private sector IT projects. Hence they face greater scrutiny from the media, the US Congress, and any number of watchdog organizations. Second, federal IT projects face challenges and risks due to increased complexity, technological uncertainty, significant resource requirements, governmental rules and regulations, and the frequent involvement of multiple stakeholders with disparate and sometimes conflicting goals. To date, much of our understanding of IT project management has been drawn from studies that have focused on the private sector (McKinsey 2012). Given the notable differences between federal and private sector IT projects and growing calls in political and media circles for the efficient utilization of tax payer contributions (Fortune 2011, P. 56), an empirical investigation of challenges in federal IT projects presents a fruitful area of research with significant implications for practitioners. Our study has two major objectives. First, we identify and conceptualize a set of salient risks in federal IT projects using a lifecycle framework. As Figure 2 indicates, we focus on the planning and execution processes within an IT project and identify three distinct types of risk—namely, complexity risk and contracting risk that arise in the planning process, and execution risk that arises in the execution process. We define each of the three risks below. As per industry standards and federal legislation (i.e., the Clinger-Cohen Act of 1996), the performance of federal IT projects is reported to the OMB using earned value management (EVM) metrics for schedule and cost performance. EVM is a project planning and control approach which compares actual accomplishment of scheduled work and associated cost against an integrated schedule and budget plan on a periodic basis. In this study, we examine the performance impact of each of three risks using a composite earned value metric, that is, schedule-cost performance index (SCPI). Second, prior research has emphasized the need for mature processes to manage IT projects; which leads to improvements in the control and predictability of project outcomes. In the context of federal IT projects, the vendor's capability to reliably deliver mission-critical IT solutions (i.e., the vendor's use of mature processes within a project) is assessed using the Capability Maturity Model Integration (CMMI) framework developed by the Software Engineering Institute (SEI). The framework consists of five levels (levels 1 – 5) which assess the evolution of a firm's processes from immature and informal to mature and formal, and define the related infrastructure necessary to support these processes at an organizational level (CMMI for Development 2010). As a signal of process excellence, CMMI level 3 represents a significant step toward process maturity, with vendor certification at this level being frequently used as a key qualifying criterion by the federal government for awarding project contracts. Given the considerable commitment of time and organizational resources required to obtain CMMI certification and the performance challenges associated with federal IT projects, we examine whether higher levels of process maturity—that is, level 3 and higher—play a significant role in mitigating the effects of risk on performance in federal IT projects. The projects for this study are drawn from a proprietary database of technology projects from Lockheed Martin, a Fortune 100 global technology firm that specializes in the development of large aerospace, defense, and security systems for the federal government (i.e., the client organization). Time-series panel data are collected across 519 quarterly time periods from 82 federal IT projects that were completed during the period 2002-2012. The firm uses a rigorous two-step procedure for collecting data on federal IT projects. In the first step, tactical and project specific details are collected on a monthly basis as part of a monthly review process. The review process is typically conducted by a panel consisting of project managers, deputy project managers, and vice presidents in the IT domain within the firm. In the second step, the monthly data are aggregated to form quarterly status reports that are used for strategic review and evaluation of project performance. In addition, these reports are used to track data for internal auditing purposes and to create lessons learned. To ensure accuracy in data collection, the data are triangulated through multiple sources (e.g., interviews with project managers, project documents, etc.). Given the time-series nature of the data, we use the generalized least squares (GLS) regression method that corrects for both panel-specific autocorrelation and heteroskedasticity in the analysis. In addition, we control for a number of factors pertaining to project characteristics (e.g., project budget, project size, project priority) in our analysis. The results provide empirical support for our arguments that each of the three types of risks—complexity, contracting, and execution risks—reduce a project's ability to meet its cost and schedule targets. More importantly, our results highlight the effect of higher CMMI levels in attenuating the negative effects of project risks on performance in federal IT projects. In addition, the attenuating effect of CMMI on the risk-performance relationship is stronger at high risk levels; at low risk levels, projects with higher maturity levels (e.g., levels 4 and 5) exhibit inferior performance on schedule and cost metrics compared to projects with lower (e.g., level 3) maturity levels. To demonstrate the economic impact of increasing process maturity levels in federal IT projects, we conduct post-hoc analysis to examine the magnitude of savings (and overruns) in project costs across different levels of CMMI and project risks. This analysis is conducted in three steps. In the first step, we estimate the dependent variable (SCPI) at low (-2 SD), average, and high levels (+2 SD) of project risks across different process maturity levels, holding all control variable values at their means. Next, based on a median project budget of $35 million in the sample, we determine the estimated cost of completion (EAC) using the formula: EAC = (ProjectBudget/SCPI) × 100. Finally, we examine the differences in EAC values across different maturity and risk levels, to determine potential savings in project budgets. The results, shown in Table 1 below, highlight the potential cost savings that may result for executing projects at higher maturity levels when project risk levels are high. Specifically, given a project budget of $35 million (based on the median project budget value in our sample), executing the project at CMMI 4 or CMMI 5 when project risks are high is associated with potential savings of $11.87 million and $8.56 million, respectively, compared to executing the project at CMMI 3. In contrast, when project risk levels are low, executing the project at CMMI 4 or CMMI 5 is associated with potential cost overruns amounting to $8.27 million and $7.26 million, respectively, compared to executing the project at CMMI 3. Findings from our study make the following important contributions to the extant literature. First, our study focuses on an important and largely understudied area of research in the OM literature—the management of public sector operations (Verma et al. 2005), and particularly, the context of federal IT projects. The second contribution of our study arises from identifying and positioning IT project risks in the context of a lifecycle framework. The importance of identifying and planning for risks has been widely discussed in the extant project management literature. Our study represents a concerted attempt to conceptualize key risks in an IT project by using the project lifecycle framework, which allows us to identify and map risks by processes associated with specific project phases. The third contribution lies in developing a nuanced understanding of the mode by which process maturity influences project performance. This is important to both theory and practice given the significant investment of resources and time that is required to acquire CMMI certification. Toward this end, our results provide the following insights to managers of federal IT projects - while the implementation of CMMI 4 relative to CMMI 3 attenuates the negative performance effects of risks in the planning process only; the implementation of CMMI 5 relative to CMMI 3 level attenuates the negative performance effects of risks in both planning and execution processes. The fourth contribution of our study is based on our results that the intrinsic benefits of CMMI implementation in federal IT projects become particularly salient at high levels of project risk; at low risk levels, the benefits of higher maturity levels (i.e., CMMI 4 and CMMI 5) on project schedule and cost metrics are inferior compared to projects with CMMI 3 maturity level. We surmise that at low levels of project risk, the improvements in project performance accruing from increased levels of process maturity, may not fully compensate the costs of implementing higher CMMI levels, thereby diminishing overall project performance. The study's final contribution arises from our use of an integrated measure of schedule-cost performance (SCPI) in the context of earned value management (EVM). While vendors working on federal IT projects are mandated to use EVM for tracking and reporting project progress as per industry standards and federal legislations, the use of EVM has also grown significantly in the private sector. Though widely used in project management, there is a dearth of studies that have used EVM for evaluating project performance. Therefore, our study provides a welcome addition to the project management literature with respect to earned value management.

We consider an inventory system under continuous review with two demand classes that are different in terms of service level required (or penalty cost incurred for backordering of demand). Prior literature has proposed the critical level rationing (CLR) policy under which the demand from the lower priority class is backordered once inventory falls below the critical level. While this reduces the penalty cost for the higher demand class, the fill rate achieved for the lower priority demand class gets compromised. In this study, we propose a new class of two‐bin (2B) policy for the problem. The proposed 2B policy assigns separate bins of inventory for the two demand classes. The demand for each class is fulfilled from its assigned bin. However, when the bin intended for the higher demand class is empty, the demand from the higher class can still be fulfilled with the inventory from the other bin. The advantage of the 2B policy is that better fill rates are achieved, especially for the lower demand class. Computational results show that the proposed policy is able to provide a much higher service level for the lower priority class demand without increasing the total cost too much and without affecting the service level for the higher priority class. When a service level constrained optimization problem is considered, the 2B policy dominates the CLR policy when the service level difference for the two classes is not too high or the service levels required for both the classes are relatively lower.

This study examines a deterministic material requirements planning (MRP) problem where lead times at subsequent ordering moments differ. Adequate replenishment methods that can cope with lead time differences are lacking because of the order crossover phenomenon, that is, replenishment orders are not received in the sequence they are ordered. This study specifies how to handle order crossovers and recalculate planned order releases after an update of gross requirements. The optimal ( s , S ) policy is based on dynamic programing. The state space is kept to a minimum due to three fundamental insights. The performance of the optimal solution approach is compared with two heuristics based on relaxations and a benchmark approach in which order crossovers are ignored. A numerical analysis reveals that average cost savings up to 25% are possible if the optimal policy is used instead of the benchmark approach. The contribution of this study is threefold: (1) it generalizes theory on MRP ordering, allowing for lead time differences and order crossovers; (2) it develops new fundamental insights and an optimal solution procedure, leading to substantial cost saving; and (3) it provides good‐performing heuristics for a general and realistic replenishment problem that can replace the current replenishment methods within MRP.

This study develops a theoretical model and then, using Canadian joint replacement surgery data, empirically tests the relationship between government policies that promote privately funded health care and patients’ waiting time in the public health care system. Two policies are tested: one policy allows opt‐out physicians to extra‐bill private patients, and the other provides public subsidies to private patients. We find that both policies are associated with shorter public waiting time, and that the subsidy policy appears to be more effective in waiting time reduction than the extra‐billing policy. Our findings are consistent with a dominant demand‐side effect in that these policies would provide patients an option, and some incentive, to opt out of the public health system, shifting the demand from the public health system to the private care market.

Profit‐maximizing firm owners who incentivize their managers with a bonus for process improvement create an intentional misalignment of their own objective and management attention. From the viewpoint of a single firm, such a local misalignment can never be profitable, but in this study we take a wider strategic perspective by investigating cost‐reducing process improvements of two firms competing in a Cournot market. We find that the use of a process improvement bonus (by firm A) can be profitable, by affecting the competitor's decision making. Informed about the reward structure at firm A, which provides an incentive for process improvement and thereby for increased production at that firm, the manager of the competing firm (B) is inclined to produce less if the owner of firm B only rewards profit. This leads to a higher profit for firm A. However, we also show that firm B's best strategy is to also offer a process improvement bonus, even if that firm is a cost laggard (with higher costs for process improvement), and that this leads to reduced profit for both firms in many situations unless one of them is sufficiently superior in its ability to improve processes. These results are robust for uncertain process improvement outcomes, multidimensional process improvement decisions, and information asymmetry in the owner–manager relationship.

We study and compare decision‐making behavior under the newsvendor and the two‐class revenue management models, in an experimental setting. We observe that, under both problems, decision makers deviate significantly from normative benchmarks. Furthermore, revenue management decisions are consistently higher compared to the newsvendor order quantities. In the face of increasing demand variability, revenue managers increase allocations; this behavior is consistent with normative patterns when the ratio of the selling prices of the two customer segments is less than 1/2, but is its exact opposite when this ratio is greater than 1/2. Newsvendors' behavior with respect to changing demand variability, on the other hand, is consistent with normative trends. We also observe that losses due to leftovers weigh more in newsvendor decisions compared to the revenue management model; we argue that overage cost is more salient in the newsvendor problem because it is perceived as a direct loss, and propose this as the driver of the differences in behavior observed under the two problems.

Seasonal demand for products is common at many companies including Kraft Foods, Case New Holland, and Elmer's Products. This study documents how these, and many other companies, experience bloated inventories as they transition from a low season to a high season and a severe drop in service levels as they transition from a high season to a low season. Kraft has termed this latter phenomenon the “landslide effect.” In this study, we present real examples of the landslide effect and attribute its root cause to a common industry practice employing forward days of coverage when setting inventory targets. While inventory textbooks and academic articles prescribe correct ways to set inventory targets, forward coverage is the dominant method employed in practice. We investigate the magnitude and drivers of the landslide effect through both an analytical model and a case study. We find that the effect increases with seasonality, lead time, and demand uncertainty and can lower service by an average of ten points at a representative company. While the logic is initially counterintuitive to many practitioners, companies can avoid the landslide effect by using demand forecasts over the preceding lead time to calculate safety stock targets.

The problem of estimating delays experienced by customers with different priorities, and the determination of the appropriate delay announcement to these customers, in a multi‐class call center with time varying parameters, abandonments, and retrials is considered. The system is approximately modeled as an M ( t )/ M / s ( t ) queue with priorities, thus ignoring some of the real features like abandonments and retrials. Two delay estimators are proposed and tested in a series of simulation experiments. Making use of actual state‐dependent waiting time data from this call center, the delay announcements from the estimated delay distributions that minimize a newsvendor‐like cost function are considered. The performance of these announcements is also compared to announcing the mean delay. We find that an Erlang distribution‐based estimator performs well for a range of different under‐announcement penalty to over‐announcement penalty ratios.

We report the results of an experimental study of route choice in congestible networks with a common origin and common destination. In one condition, in each round of play network users independently committed themselves at the origin to a three‐segment route; in the other condition, they chose route segments sequentially at each network junction upon receiving en route information. At the end of each round, players received ex‐post complete information about the distribution of the route choices. Although the complexity of the network defies analysis by common users, traffic patterns in both conditions converged rapidly to the equilibrium solution. We account for the observed results by a Markov adaptive learning model postulating regret minimization and inertia. We find that subjects' learning behavior was similar across conditions, except that they exhibited more inertia in the condition with en route information.