Organizing the Forecasting Department

Throughout the years that I have been giving presentations on demand forecasting I've gotten a lot of questions asking me: How does one organize a forecasting department? I usually discuss this aspect as part of my tutorial on designing a forecasting process for The Institute of Business Forecasters' annual sessions - because the success of the process ultimately hinges on the performance of the forecasting department and its staff. Unfortunately, however, I don't have a standard pat answer for how to organize a department because there is no one way that works best. A combination of skills, organization, and corporate culture form the cocktail that drives the success of any forecasting department. WHAT FORECASTING SKILLS ARE NEEDED? First and foremost the success of a forecast department depends on its people. I frequently advise people not to staff their department only with PhDs from quantitative disciplines, because that will likely be a recipe for disaster. Quantitative skills, while naturally necessary (since forecasting is about generating numeric projections), are not the only skills needed within a department. Interpersonal skills are equally as important. People in the department must be able to communicate with others in order to gather the market intelligence needed in developing and adjusting baseline forecasts, incorporating the views of others, identifying factors that will impact demand, and driving to a consensus-based forecast. While no one person in a forecasting department needs to have all the skills, the composite group needs to have a good mix of the following skill sets: * Quantitative Skills: These are needed because forecasting often involves the use of statistical forecasting methods and algorithms. Many times errors may be made and will go undetected if someone is not savvy enough about the forecasting methods to determine that a bad number was generated. For example, if an abnormal period of historical demand is used as an input to a forecasting algorithm, a forecast might be way off and this could go unnoticed if not looked at by someone that understands numbers and their implications. * Computer Skills: Since demand forecasts often need to be done at the Stock Keeping Unit (SKU) level by stocking location for multiple periods of time, the number of forecast entities that need to be developed can get quite large - in some cases in the tens or hundreds of thousands of entities or more. Automation is required, and this requires the skills needed to run computer software applications. * Oral Communication Skills: This is a subtle and often overlooked set of skills, but an extremely important set of skills to have in a forecasting department. Good forecasters need to be able to solicit any and all information that could impact demand. To do this involves communicating often and well with all the stakeholders in a company that understand and could impact future demand. For example, when I was a forecaster I kept in constant communication with sales and marketing management to keep my fingers on the pulse of the business. This allowed me to identify salient demand shifts before or as they occurred. * Understanding of the Business: This is another extremely important set of skills to have because it helps to identify demand variations. Forecasting is all about understanding demand variations because if they were no variations in demand there would be no reason to have to forecast. Demand would be flat! An understanding of the business is needed to help identify customer and sales channel trends that are evolving, as well as the reasons behind seasonal demand patterns and business cycle shifts. * Process Management Skills: This set of skills is needed to help ensure that the process is successful in getting to a one-number forecast, a best business practice. This entails getting a cross-functional team to come to a consensus on the forecasts. …