Industrial Management

MAR-APR 2014

Issue link: https://industrialmanagement.epubxp.com/i/285723

Contents of this Issue

Navigation

Page 27 of 31

28 Industrial Management the conditions that generated past data are indistinguishable from the condi- tions of the future. The recognition that forecasting techniques operate on the data generated by historical events leads to the identification of the following steps in the forecasting process: 1. Problem definition and data collection 2. Data manipulation and cleaning 3. Model selection and evaluation 4. Model implementation (actual forecasts) 5. Forecast evaluation 6. Forecast presentation 7. Tracking results It should be emphasized that management ability and common sense must be involved in the forecasting process. The forecaster should be thought of as an advisor to the manager rather than as a monitor of an automatic decision-making device. The quantitative techniques in the forecasting process must be seen as what they really are, namely, tools to be used by the manager in arriving at better decisions. In addition, a more realistic attitude can improve the usefulness and utility of forecasting. Forecasting should not be viewed as a substitute for prophecy, but rather as the best way of identifying and extrapolating established patterns or relationships in order to forecast. With that being said, if management of the forecasting process is to be conducted properly, several key questions always should be raised, such as the following: • Why is a forecast needed? • Who will use the forecast? • What are the specific requirements? • What level of detail or aggregation is required? • What is the proper time horizon? • What data are available? • Will the data be sufficient to generate the needed forecast? • What will the forecast cost? • How accurate can we expect the forecast to be? • Will the forecast be made in time to help the decision-making process? • Does the forecaster clearly under- stand how the forecast will be used in the organization? • What could change the forecast? Biases in forecasting The chance of a forecast to be on target will increase if there is a process in place that systematically identifies and mitigates biases, as K.B. Kahn wrote in "Identifying the Biases in New Product Forecasting," from the Journal of Business Forecasting. It is unlikely that the forecast of a product's sales will be completely free from all biases because, after all, forecasts are based on judgment. However, understanding the most prevalent biases, combined with making the process as transparent as is feasible, can help make better forecasting decisions. Bias comes in two forms: uninten- tional and intentional. Unintentional bias is an unforeseen and uncontrol- lable error. It is caused by process deficiencies and a lack of information or management experience in building forecasts. An example of unintentional bias is the deviation between the actual sales of a new product and what was expected. Intentional bias, on the other band, is error that is introduced purposely, though sometimes not consciously. Intentional bias stems from personal motivations and company politics. It is caused by targets and incen- tives imposed on those who build the forecast. These incentives and targets politicize the forecasting process, resulting in reduced forecast accuracy. The politics of forecasting are encountered whenever a forecaster is biased to forecast volumes that support someone else's expectations. This intentional biasing is a major roadblock to obtaining accurate forecasts. For example, a salesperson is given an incentive plan that rewards exceeding targets instead of rewarding forecast accuracy. Typically, a salesperson who hits 80 percent of the target gets no bonus, whereas a salesperson who hits 120 percent of target receives a large bonus. This incentive heavily biases the salesperson to forecast lower sales totals in order to maximize bonus payout. Suppliers trying to supply these salespeople will face problems. Sales forecasts that are too high could result in excess inventories, while forecasts that are too low could generate higher costs when shipments need to be expedited to meet unexpected demand. Marketing managers must justify their allocation of the marketing budget based on projected product and customer growth. This biases them to predict higher sales in order to keep their marketing budgets from shrinking. Top-level growth targets also will bias marketing personnel to predict that consumers will accept new products or features readily, and that the overall market will grow. Research and development can be biased to justify increased development budgets. Supply chain executives can add bias to handle their inventory management and capacity utilization targets. Financial analysts can add bias to the forecast through their efforts to manage shareholder expectations. Purchasing agents can add bias to the forecast in an effort to negotiate more favorable pricing, as Tom Wallace reported in "Forecasting and Sales and Operations Planning: Synergy in Action" in the Journal of Business Forecasting. Intentional bias also is caused by the budgeting process itself. The budgeting process is an exercise in setting revenue targets based on market conditions at a specific point in time. As the recent recession demonstrated, market condi- tions can change dramatically and quickly after the budget has been locked in. However, each time the forecast is updated, the forecasters are biased to restate the budget. Forecasters were slow to forecast the decrease in demand during the recession because their initial bias of building a budget during good times had not gone away. These incentives and targets politicize the forecasting process, resulting in reduced forecast accuracy. IM MarApr 2014.indd 28 3/24/14 11:02 AM

Articles in this issue

Archives of this issue

view archives of Industrial Management - MAR-APR 2014
loading...
Industrial Management
Remember me