Supply Chain Digest editorial staff

The News: The Journal of Business Forecasting releases its annual benchmark results

The Impact: Opportunity for companies to see how they compare with others in terms of forecasting and demand planning/management practices.

The Story: The Journal of Business Forecasting recently released its fourth annual benchmarking report, compiled by Chaman Jain of St. John’s University. The results are from attendees at conferences of the Institute of Business Forecasting , publisher of the journal. About two-thirds of respondents came from companies with over $500 million in revenue.

Key findings include:

• Management Support: 43% of respondents said management is highly supportive of the forecasting process, and 51% said it was somewhat supportive.
• Reporting Structure: 26% of the companies said forecasting reported into Operations/Manufacturing, and 12% into the Logistics group. Combining these two, 38% say forecasting is part of a supply-chain related group. 8% said they had a separate forecasting function/department. 17% and 13% said they reported to Sales and Marketing, respectively. 12% reported into Strategic Planning, while only 5% said they reported into Finance – a steep drop from just a few years ago.
• Internal Disconnects: 64% said there were goal conflicts between functions related to the forecasting process.
• One View of the Truth: 44% of respondents said their companies use multiple forecasts, versus a single consensus forecast. Supply Chain Digest believes this number is actually low compared to the reality.
• Forecast Horizon: The plurality of companies said they operate on a one-year forecasting model (46%). 22% reported they operated with a greater than 12-month horizon. 17% indicated they forecast one quarter ahead, and 14% less than a quarter.
• Production Freeze: 52% of companies said they lock production schedules one month out. 13% said two months out, and 19% said three months out, with smaller percentages for even longer horizons. It is not clear how the survey handled make-to-order or other companies that do not lock production at all.
• Consensus Forecast Meeting: A “poor man’s” version of Sales and Operations Planning (S&OP), 74% of respondents said they at least had a cross-functional meeting to agree on a consensus forecast.
• S&OP: 60% of respondents said they had a formal S&OP process in place. This number is 10-15 percentage point slower than is often found in such studies. However, the percentage of how effective the S&OP process is generally well-under the number that say they have a formal process.
• Collaborative Planning, Forecasting and Replenishment: 26% of respondents said they were involved in CPFR. Unsurprisingly, consumer goods companies were the most active.

Jain also benchmarked the forecasting models companies are using. Various time-series models were the most prevalent, with 68% of respondents indicating they used some form of time-series as their primary model. 20% said cause-and-effect models predominated, and 12% “judgmental,” which as the author points out does not mean “seat of the pants,” but rather structured techniques used when there is little historical data. An analog forecast is one such approach, for example, using the history of MP3 players to forecast a new-to-market handheld electronics. device.

Of course, many companies in practice use a combination of these techniques depending on the product or where it is in its lifecycle.

Among time series modelers, the specific technique the predominated broke out like this:

• Averages/simple trending: 61%
• Exponential smoothing: 29%
• Box Jenkins: 7%
• Decomposition: 3%

Among cause-and-effect modelers, the specific techniques used were:

• Regression: 79%
• Econometric: 16%
• Neural Networks: 5%

Do these benchmark numbers seem about right to you? If not, why not? What would you like to see in forecasting benchmarking? Let us know your thoughts.

Article key words: Forecasting, demand planning