Given the critical role of the forecast it amazes that today supply chain organizations still routinely rank inaccurate forecasts as a major obstacle in meeting their supply chain goals. Obtaining a consistently good forecast accuracy continues to be a never-ending struggle that requires skills in math, logic, communication, collaboration and knowledge of the products, customers, markets and industries for which you create the forecasts. It doesn’t help that finding and retaining qualified planning talent increases the pressure on existing team members as open positions go unfilled. This post provides five tips for improving a demand planner’s ability to develop more accurate forecasts.
Tip #1 — Deploy a Position Rotation Program: To be truly effective demand planners need to understand how to minimize forecast error and how their forecasts are used. Ask yourself, “is the overall objective to create the most accurate forecasts or is it to create forecasts that minimize costs and maximize company revenue and profit?” Gaining hands-on experience in using forecasts to purchase product and plan inventory, production and distribution will provide demand planners a unique understanding of how to make their forecasts more useful.
Tip #2 – Certification / System Training: There is a high correlation between skill certification and system training and the ability to create accurate forecasts. Many variables come into play that affect the ability to create accurate forecasts. Some of these variables are uncontrollable others are partially controlled. Demand planner training is one of the few variables you can control.
Tip #3 — Switch to Using Order Data: Using quantity shipped per period can significantly reduce forecast accuracy because this data may not accurately reflect actual customer needs/demand. Shipment data often encompasses misleading information due to inventory shortages, distribution mistakes and delays, non-repeating calendar effects and other variables that negatively effect On-Time Perfect Order performance. If these non-repeatable errors and events are part of the data used to generate forecasts, your accuracy will be impacted.
Tip #4 — The Importance of Data Scrubbing: Forecasts are only as good as the data used to generate them. Even if you use actual demand history there will be outlying data due to non-repeatable promotions, competitor actions, unplanned events, weather effects, social media impacts, etc. that needs to be addressed to create a “corrected” historical file. Many demand planning processes do not build enough time in to the periodic schedule to adequately scrub, or correct, the data used to create forecasts. Forecast accuracy can be significantly improved by allowing a bit more time up front to correct the historical data. Advanced planning systems can automatically filter outlying data points, but it is better to do this prior to forecast creation.
Tip #5 — Use Multivariate Pareto (ABC) Analysis: Pareto analysis, also known as ABC classification, is a useful method for classifying forecasted items according to their relative importance to the company. For example, if the Pareto Analysis is by volume then items that represent 80% of total volume are categorized as type “A”, items representing the next 15% of total volume are categorized as type “B” and slower-moving items representing the last 5% of total volume are categorized as type “C”. A rich forecasted item profile can be developed by creating a Pareto segmentation on multiple additional variables such as revenue, profit, production resource, geographic region, business segment, etc. This profile allows a demand planner to analyze forecast accuracy from a number of different ways to help focus their time and efforts on improving the forecasts of items that will make the biggest impact.
Better forecasting is one of the few “power-ups” available that can drive both top line revenue and bottom line profitability. Creating accurate and useful forecasts requires a blending of art and science and a demand planner’s effectiveness depends heavily on a well thought out process, their access to advanced technology and their level of experience and training.
What do you think? What actions is your company taking to improve its ability to predict demand?