Overcoming Forecast Error with Real-time Forecasting

Working to improve demand visibility at your company but stumped on how to improve demand accuracy? Here are some tips to get you started

Working to improve demand visibility at your company but stumped on how to improve demand accuracy? Here are some tips to get you started.

Leading consumer goods manufacturers have seen the AMR Research reports advocating the Demand-Driven Supply Network (DDSN) and, in many cases, are leveraging these concepts to launch initiatives focused on improving demand visibility and accuracy. While the DDSN framework acknowledges that demand accuracy is the most important supply chain metric, it does not provide detailed guidelines for improving it. High forecast error rates can lead companies to produce unneeded inventory, deliver incomplete orders and have difficulty maintaining stable production schedules.

Given the enterprise resource planning (ERP) and planning-related efficiency gains achieved over the past decade, it may be difficult to believe that leading consumer goods companies continue to operate with forecast error rates around 50 percent by item/distribution center/week. Let's review why this happening.

First, although the sales and operations planning (S&OP) processes consolidate relevant data from across the organization, the resulting plan remains static until the next planning cycle, typically a month away.

Second, because the S&OP plan is focused on long-term demand planning and not refreshed with daily transactions, data validity begins to almost immediately deteriorate, rendering it ineffective for operational purposes.

Third, the S&OP planning process focuses on generating product family level plans; however, more detailed stock-keeping unit (SKU)-level data is required to meet the demands of today's consumer-driven market.

Finally, collaborative initiatives such as collaborative planning, forecasting and replenishment (CPFR) and vendor-managed inventory (VMI) have not delivered the results that were originally expected.

Traditional Systems Not Responding to Real-time Opportunities

Planning systems provide the technology foundation for S&OP processes for many leading organizations. They deliver reasonable long-range forecasts by combining historical sales with statistical predictions, typically on a monthly basis. As business processes were developed to support the implementation of these systems, many leading consumer goods companies established best practices that institutionalized cross-functional discipline throughout their own organizations.

Demand planning applications are not designed to respond to the real-time needs of today's fast-paced market. They are neither precise nor responsive enough to drive programs such as the perfect order, just-in-time or lean. Implementing these market-driven programs requires new business processes and the integration of real-time data for informed and rapid decision making. Real-time forecasting systems are designed to use detailed order information to accurately predict daily demand over the next 30 to 60 days, thereby generating the best possible forecast each day.

Real-time Forecasting in Action

Real-time forecasting solutions are relatively new, but there are a few companies that are pioneering the processes and systems. For instance, Campbell Soup Co. and Ventura Foods have leveraged real-time forecasting to substantially lower forecast error rates, which has led to favorable inventory reductions and improved customer service.

A few years ago, Campbell initiated an enterprise-wide innovation program that focused on more quickly introducing new products to the market. While this effort has improved top-line revenue, it also placed additional pressure on already strained demand planning and inventory management systems. The company launched a program to implement real-time forecasting across its North American operations. As a result, Campbell is now able to get new products to the market faster than ever before. Real-time forecasts are generated each day and have helped to reduce finished goods inventory by more than a week, stabilize manufacturing schedules and decrease logistics costs.

Best known for its Hidden Valley Ranch dressings, Ventura Foods is a leading provider of edible oils, shortenings, dressings, margarine and sauces. Ventura has historically struggled with balancing consumer demand, coordinating a dozen manufacturing plants and negotiating with a volatile commodities market. A couple of years ago, it launched an initiative to improve customer service, stabilize manufacturing schedules and decrease inventory buffers. By improving planning systems and implementing real-time forecasting processes and systems, Ventura's increased visibility has resulted in better asset utilization and reduced risk of raw material cost fluctuations.

Charting a Better Forecast for the Future

Monthly S&OP processes and systems are no longer sufficient to meet today's demands. While AMR's Demand-Driven Supply Network provides a technology framework worthy of consideration, each organization must evaluate its goals relative to opportunities to generate short-term improvements.

Common issues for consumer products manufacturers include excess inventory, faltering customer service and lack of manufacturing predictability; each of these issues can be traced to forecast error. If your organization is faced with these issues and would like to take action, you may want to begin by determining the role and the extent of today's forecast error. If you are uncertain as to the error rate or the impact that it is having on your business, a solution provider should be able to help you quantify the problem. When you are ready to work with solution providers, work only with those vendors that know the industry, understand your business and have the intellect to determine scope of the solution.

In addition to quantifying the issue, you should also better understand how the S&OP process works at corporate planning sessions, and how these plans are being implemented by operations managers throughout the business. Identify the S&OP participants, gauge the accuracy of the monthly plan and determine how effectively these managers are driving results back into their own organizations. Remember to talk with operations managers to understand how well they receive and trust the information that they are given to do their daily jobs. When you are ready to work with solution providers, work only with those vendors that have experience with these processes, understand the impact of incorrect information and are realistic about the scope of effort required to drive change.

Prior to choosing a solutions provider, ask for an illustration of how its system would have generated results better than those generated by your existing systems. Work only with those vendors that are willing to benchmark their systems against your current environment.

About the Author: As President of Biddle Associates, Jane Biddle leverages her 25+ years of manufacturing and supply chain industry technology experience to provide advisory and consulting services to the firm's clients. Biddle works with technology executives to create product and market strategies; and with manufacturing executives to develop global IT strategies. Prior to Biddle Associates, Biddle was chief commercial officer for INTTRA. Previous positions include vice president of Marketing for G-Log and RockPort Trade Systems, Industry director for SAP and Research director for Benchmarking Partners, MRP/ERP Product manager for Digital Equipment, and Manufacturing consultant for Hewlett-Packard.

For more information on the supply chain execution market, read "Demand Management," part of the Analyst Corner Series, in the February/March 2005 issue ofSupply & Demand Chain Executive magazine.

Companies in this article