Larry P. Ritzman

Integrating judgmental and quantitative forecasts: methodologies for pooling marketing and operations information

Accurate forecasting has become a challenge for companies operating in todays business environment, characterized by high uncertainty and short response times. Rapid technological innovations and e‐commerce have created an environment where historical data are often of limited value in predicting the future. In business organizations, the marketing function typically generates sales forecasts based on judgmental methods that rely heavily on subjective assessments and “soft” information, while operations rely more on quantitative data. Forecast generation rarely involves the pooling of information from these two functions. Increasingly, successful forecasting warrants the use of composite methodologies that incorporate a range of information from traditional quantitative computations usually used by operations, to marketings judgmental assessments of markets. The purpose of this paper is to develop a framework for the integration of marketings judgmental forecasts with traditional quantitative forecasting methods. Four integration methodologies are presented and evaluated relative to their appropriateness in combining forecasts within an organizational context. Our assessment considers human factors such as ownership, and the location of final forecast generation within the organization. Although each methodology has its strengths and weaknesses, not every methodology is appropriate for every organizational context.

Intended and Achieved Competitive Priorities: Measures, Frequencies, and Financial Impact

Using data collected from managers in a sample of manufacturing plants, parallel sets of variables representing intended performance and achieved performance are factor analyzed to see if they break out in a similar way. For each competitive priority identified as a factor (latent construct), a frequency table is constructed showing how intended and actual competitive priorities match up in practice. Finally, an overall performance measure is overlaid onto the frequency matrix showing how competitive priorities ╌ both intended and realized ╌ affect a firm’s bottom-line financial performance.

Judgmental Adjustment of Statistical Forecasts

Judgmental and statistical forecasts can each bring advantages to the forecasting process. One way forecasters can integrate these methods is to adjust statistical forecasts based on judgment. However, judgmental adjustments can bias forecasts and harm accuracy. Forecasters should consider six principles in deciding when and how to use judgment in adjusting statistical forecasts: (1) Adjust statistical forecasts if there is important domain knowledge; (2) adjust statistical forecasts in situations with a high degree of uncertainty; (3) adjust statistical forecasts when there are known changes in the environment; (4) structure the judgmental adjustment process; (5) document all judgmental adjustments made and periodically relate to forecast accuracy; (6) consider mechanically integrating judgmental and statistical forecasts over adjusting.

The relative significance of forecast errors in multistage manufacturing

Abstract It is intuitive that controlling forecast errors should result in better customer service and lower inventories. But forecast errors come from various sources. Multistage manufacturing experiences supply and lead-time uncertainties as well as demand uncertainties, and manufacturing has more than one lever to pull when addressing these uncertainties. The relationship of forecast errors to manufacturing performance is not clear. Furthermore, the pursuit of forecast accuracy may not be the best use of managerial resources. In this study, using a many-factored manufacturing simulation, we examine two components of forecast errors, the mix of special and standard products, lot-sizing, and buffering policies as they affect inventories and customer service. Although most conclusions are situation dependent, reducing forecast bias is shown to be much preferred to reducing forecast variability, bias management is more important to on-time delivery than to inventory reduction, and the value of such reductions is particularly important in situations where there are large lot-sizes and small buffers.

Uncertainty Reduction Approaches, Uncertainty Coping Approaches, and Process Performance in Financial Services

Developing a better understanding of the impact of uncertainty on process performance has been recognized as an important research opportunity in service design (Hill, et al., 2002). Within this general research stream, our study focuses on the question of what managers can do to most effectively address operational uncertainty and mitigate its negative effects. To begin to address this question, we report on an exploratory study using a sample of professionals in the financial-services industry who acted as informants on 108 financial-services processes. These professionals were sampled from a population of graduates of a university in the northeastern region of the United States who were employed in the financial-services industry. Based on these processes, we empirically examine the relationship between responses to operational uncertainty and process performance after controlling for customer mix, other uncertainty sources, and process type characteristics. Our findings suggest that process improvement—an uncertainty reduction approach related to the internal functioning of the process—as well as several uncertainty coping approaches are associated with better performing processes. However, uncertainty reduction approaches related to customer involvement with, and demands on, the process are not associated with better performing processes. We discuss the implications of our findings for determining what actions managers can take to reduce the negative performance effects of operational uncertainty and how managers can decide which of these actions to take. We conclude with a discussion of the limitations of our study.

A cyclical scheduling heuristic for lot sizing with capacity constraints

This paper proposes a heuristic procedure to solve cyclical scheduling problems. Lot sizes for two or more items must be chosen simultaneously for several periods into the future, while recognizing capacity constraints at several work centres. Solving this problem is important in its own right. It also has transfer value to even more complicated situations, such as when the items have parent-component relationships. The proposed heuristic, though in the spirit of Mamies formulation, has fewer production sequences for each item. The simplified formulation can be solved with a general linear programming code. Another advantage is its use of fixed ordering intervals, which seems to be preferred in practice. Experimental results suggest that the heuristic gives near optimal solutions and requires less computational time. Of particular interest is the heuristics ability to satisfy capacity constraints when selecting item lot sizes.

Bringing judgment into combination forecasts

Abstract This research investigates the benefits in forecast accuracy by combining judgmental forecasts with those generated by statistical models. Our study differs from prior research efforts in this area along two important dimensions. First, two different types of judgmental forecasts are evaluated for combination with statistical forecasts — one based on contextual knowledge and one based on technical knowledge. Contextual knowledge is information gained through experience on the job with the specific time series and products being forecasted. Technical knowledge is information gained from education on formal forecasting models and data analysis. Second, we investigate the conditions under which adding judgment to combination forecasts helps the most. Specifically, we test the improvement as a function of time series variability. Our results show that judgmental forecasts based on contextual knowledge, rather than technical knowledge, are the better input into combination forecasts. Bringing judgmental forecasts based on contextual knowledge into combination forecast improves forecast accuracy over the individual statistical and judgmental forecasts. However, the benefit attained from including contextual knowledge in the combination depends on the amount of inherent variability in the time series being forecast. More contextual knowledge is needed for combination forecasts if a time series has more data variability. If the amount of variability is low, less emphasis should be given to contextual knowledge when making combination forecasts. In general, our findings suggest a linear relationship between the amount of contextual knowledge needed and data variability.

An empirical investigation of different strategies for material requirements planning

Abstract Master production scheduling, component lot sizing, and capacity requirements planning represent three important modules of material requirements planning (MRP) systems. Coordinating these three modules has been largely dependent on managerial judgment and experience. In this research, five different strategies for integrating these modules are empirically investigated. These strategies differ mainly in the extent to which the modules are coordinated. The impact of the shop, product, component and cost characteristics is measured by varying ten experimental factors. The research findings so obtained provide several guidelines on the effectiveness of each strategy in different environmental settings.

A heuristic algorithm for capacity sensitive requirements planning

Abstract Available lot sizing rules for use in MRP (Material Requirements Planning) systems ignore capacity limitations at various work centers when sizing future orders. Planned order releases are instead determined by the tradeoff only between the items set up and inventory holding costs. This limitation can cause unanticipated overloads and underloads at the various work centers, along with higher inventories, poorer customer service, and excessive overtime. This research explores one way to make MRP systems more sensitive to capacity limitations at the time of each regeneration run. A relatively simple heuristic algorithm is designed for this purpose. The procedure is applied to those planned order releases that standard MRP logic identifies as mature for release. The lot sizes for a small percentage of these items are increased or decreased so as to have the greatest impact in smoothing capacity requirements at the various work centers in the system. This algorithm for better integrating material requirements plans and capacity requirements plans is tested with a large scale simulator in a variety of manufacturing environments. This simulator has subsequently undergone extensive tests, including its successful validation with actual data at a large plant of major corporations. Simulation results show that the algorithms modest extension to MRP logic significantly helps overall performance, particularly with customer service. For a wide range of test environments, past due orders were reduced by more than 30% when the algorithm was used. Inventory levels and capacity problems also improved. Not surprisingly, the algorithm helps the most (compared to not using it at all as an MRP enhancement) in environments in which short-term bottlenecks are most severe. Large lot sizes and tight shop capacities are characteristic of these environments. The algorithm works the best when forecast errors are not excessive and the master schedule is not too “nervous.” This proposed procedure is but one step toward making MRP more capacity sensitive. The widely heralded concept of “closed-loop” MRP means that inventory analysts must change or “fix up” parts of the computer generated material requirements plan. What has been missing is a tool for identifying the unrealistic parts of the plan. Our algorithm helps formalize this identification process and singles out a few planned order releases each week. This information comes to the analysts attention as part of the usual action notices. These pointers to capacity problems go well beyond capacity requirements planning (CRP) and would be impossible without computer assistance. Our study produced two other findings. First, short-term bottlenecks occur even when the master production schedule is leveled. The culprits are the lot sizing choices for items at lower levels in the bills of material. “Rough-cut” capacity planning, such as resource requirements planning, therefore is not a sufficient tool for leveling capacity requirements. It must be supplemented by a way to smooth bottlenecks otherwise caused by shop orders for intermediate items. Second, the disruptive effect of large lot sizes is apparent, both in terms of higher inventories and worse customer service. Large lot sizes not only inflate inventories, but paradoxically hurt customer service because they create more capacity bottlenecks. The only reason why management should prefer large lot sizes is if set-up times are substantial and cannot be efficiently reduced. This finding is very much in step with the current interest in just-in-time (JIT) systems.

The Value of Outsourcing: A Field Study

Abstract This article examines the effects of information systems outsourcing on the business processes of organizations. Rather than simply comparing outsourcing and not outsourcing, the study also addresses a third and increasingly common strategy, that of using software purchased “off-the-shelf.” An extensive survey was distributed to business process managers over a cross-section of financial services processes and companies. Results show that outsourcing information systems can create lower overall process costs and may lead to superior overall process performance compared to processes that used software purchased off-the-shelf. Further, information systems built in house lead to superior overall process performance compared to processes that used software purchased off-the-shelf. These results should assist business managers in gauging the possible effects of outsourcing information systems (or not) on their core processes.