Blog

Demand Forecasting Excellence: Foundation for a Highly Efficient Supply Chain

September 28, 2012


demand-forecasting-116-x-250-(2).jpgEstimating future demand is one of the most fundamentally valuable, but frustratingly difficult, challenges in supply chain optimization. By “Demand Forecasting,” I mean the effort to increase forecast accuracy and customer service levels through better perceiving, predicting and shaping the full range of factors that determine how well your product portfolio satisfies market needs. No other aspect of supply chain optimization has greater impact on profitability.

 
We recently published a free white paper on the four key “pillars” that support effective demand planning, higher service levels and lower inventory cost.
 
  1. Go beyond simple forecasting: Increase forecast accuracy by better perceiving and predicting the full range of factors that influence demand signals. Draw from a range of statistical methods to create a forecasting model that best fits the complete situation, including a baseline forecast plus seasonal, promotional, and other components. Use attribute-based models for products with little or no reference history. Adopt a demand planning strategy that automates model selection, automatically monitors the fit against actual data, and recommends more accurate alternative models as needed throughout the product life cycle. 
  2. Beat the “Devil in the Details” using a demand disaggregation: Break down forecasts at a higher level of aggregation into detailed plans covering product components such as style, color, size, sales channel, customer, region, etc. Employ a pyramid of multiple faces, in order to aggregate different characteristics important to different stakeholders. While one face of the pyramid rolls up to an aggregate forecast by customer, other faces may by organized by geography, or product family, and so on. 
  3. Take planner productivity to the next level: Adopt a “set-and-forget…until alerted” philosophy around forecasting. A good demand planning system can automatically compare an existing demand curve to the actual demand signal as sales data becomes available. Set thresholds such that real-time email alerts to focus planners’ attention first on critical SKUs that deviate from expectations and require further analysis or intervention.
  4. Make collaboration a core demand planning competency: Use collaboration techniques such as CPFR and VMI to organize and share information with customers, partners, and internal stakeholders via a secure, Internet-based system. Capture the demand signal as close to the end-user as possible through fast feedback loops. Base sales and operations planning (S&OP) sessions on statistically sound forecasts and fresh demand data to take the emotion out of negotiations. Focusing on a few key aspects of demand forecasting helps your organization better perceive and predict the full range of factors that influence demand signals, break down forecasts into detailed plans at the attribute level, adopt a “set and forget…until alerted” planning philosophy, capture the demand signal as close to the end user as possible, and base sales and operations planning (S&OP) decisions on solid statistical science.
“Four Pillars of Demand Planning Excellence” is available now as a free download. (Registration required for new visitors.)


John Brewer is Director at Logility. He can be reached at blog@logility.com.

Tags: Demand Planning,S&OP

Logility Blog PostsBlog RSS

Blog Subscription

Supply Chain Brief

Featured Video