Supply Chain Best Practices Consortium Sales Forecasting Tools and Methods Executive Seminar Supply Chain Leadership Forum Track A, Session 4 September 11, 2007 1
Session Scope This session covers best practices and relevant data for sales forecasting. We will discuss the following subject areas: Forecasting Parameters Forecasting Staffing Forecasting Systems/Applications Used Forecasting Techniques Used Forecasting Ownership and Degree of Internal Collaboration Collaborative Forecasting with External Organizations Improvement of Current Forecasting Process 2
Forecasting History, Horizons, Intervals and Entities: Overall, Across All industries: The most commonly used forecasting horizon is 1 year. The most commonly used forecasting interval is the month. The average history used is 2 years. The average application forecasts: 2.8 product hierarchy levels, 2.1 geographic market hierarchy levels, and 2.8 customer hierarchy levels. 3
Forecasting History, Horizons, Intervals and Entities (continued): By Industry: CPG Manufacturing: 40% use horizons of 2 years and over. All use horizons of 1 year with monthly intervals. The sales history used is up to 2 years. Retail: 50% of the organizations forecast down to the day level. 50% of the organizations use monthly horizons with weekly intervals. The sales history used is 2 to 3 years. 4
Forecasting Staffing: Forecasting Staff Minimum Average Maximum CPG Manufacturing 0 8.0 40 Retail 3 5.0 8 Overall 0 7.1 40 Forecasting staffs differ widely in size, as well as in background. Purchasing / Replenishm ent 10% Marketing 20% Inventory 10% Skills Required Sales 20% Prior Forecasting 40% 5
Forecasting Software: Overall, Across Industries: All of the automated systems allow multiple users to edit forecasts simultaneously. The average cost for a forecasting system implementation is $425,000. The majority of the respondents use automated systems. The annual cost of maintaining the system ranges from 11-15%. 6
Forecasting Software (continued): By Industry: CPG Manufacturing: 60% of respondents use automated systems. The software used is predominantly developed inhouse, using a VB/SQL platform. Retail: Virtually all respondents use automated systems. The software used is predominantly purchased. 7
Forecasting Techniques: Overall, Across Industries: The forecasting software offers 4 to 10 techniques, out of which, 75% are used. Virtually all routine demand forecasts are in units, while 2/3 are also in dollars. 45% of the software adjustment factors CPG Manufacturing: are used, Breakdown of Historical Sales particularly returns and out-of-stock periods. Seasonal 8% The most used algorithms are moving average and Trend Repeat 9% History exponential smoothing. 71% Noise 12% 8
By Industry: CPG Manufacturing: 40% take returns into account. Retail: 33% take out-of-stock periods into account. CPG Manufacturing: Breakdown of Historical Sales Retail: Breakdown of Historical Sales Seasonal 8% Noise 12% Noise 13% Trend 9% Repeat History 71% Seasonal 20% Repeat History 57% Trend 10% 9
Forecasting Ownership and Internal Collaboration: Overall, across industries: 50% have little or no internal collaboration. 30% have independent processes which are then reconciled by negotiation. 20% have consensus processes. 10
By Industry: Most Significant Forecast Influences CPG Manufacturing: The dominant influences are Sales, Management, the Sales Force, and Marketing. Retail: The dominant influence is the Purchasing and Replenishment Team. R&D / Finance Engineering 6% 6% Manufacturing 6% Marketing / Product Management 17% Sales Force 17% Sales Management 27% Purchasing / Replenishment 21% 11
Collaborative Forecasting with External Organizations: 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Degree of Upstream Collaboration with VENDORS 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Degree of Downstream Collaboration with CUSTOMERS 12 Upcoming projects/large deals Sales History Sales Forecast Promotion and advertising Near real time sales data (e.g. POS Upcoming projects/large deals Sales History Near real time sales data (e.g. POS Promotion and advertising Sales Forecast Overall, across industries: The majority of collaboration with external groups is downstream to customers.
100% Approaches to improving forecasts cover a wide range of technological, process, and participatory initiatives and vary almost as widely as forecasting staffs. 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Perceived Importance of Factors to Improve Current Forecasting Process 13 Consulting Assistance Software Upgrade Forecasting Technique Sophistication Budget Marketing / Product Development Collaboration ERP Integration Online Access Staffing Customer Collaboration Finance / Planning Collaboration Senior Management Involvement Vendor / Supplier Collaboration Process Documentation Sales Force Collaboration Supply Chain / Logistics Collaboration Sales Management Collaboration Manufacturing Collaboration Improvement of the Current Forecasting Process:
Improvement of the Current Forecasting Process (continued): and the complimentary avenues are both important as well as equally diverse. Capacity Guarantees from Vendors / Expedited Suppliers Transportation 6% Options 11% Sourcing Options for Key Products 11% Avenues to Complement Forecasting Process Near-real-time Visibility of Actual Demand 22% Increased safety stock 17% Shorter Lead Times for Hard-toforecast Products 33% 14
Potential Group Discussion Points Are the financial and customer service price tags for inaccurate forecasts well understood by senior management? How do you prioritize initiatives to improve forecasting? What forecast horizon and interval provide the best results for your company? Do you understand why? How effectively do you share forecast information with external organizations. (i.e., just because they get it, do they get it? ) At what level of aggregation should forecast accuracy be reported to be most effective? 15
Important Takeaways While there are many Sales Forecasting takeaways in Benchmarking & Best Practices, some of the more important are: Determination of the best forecast horizon and interval is a critical part of generating meaningful forecasts. Forecasting is much, much more than software and algorithms. Although primarily after-the-fact, the avenues for hedging forecast accuracy are very important aspects of forecasting. 16
Benchmarking & Best Practices References Information on Sales Forecasting and related topics can be found in the following references: Topic: Forecasting Sections: o Process and Tools o Techniques o Process Inputs o Systems and Models o Accuracy o Outcomes 17