2015 Forecasting Benchmark Study
It is not a pretty picture. Item proliferation is rampant and getting worse, with growth through innovation strategies driving complexity but not sales. The days of relying on historical performance and applying rules of thumb are over, replaced by the use of daily data reflecting current market conditions, automation and advanced pattern recognition algorithms.
Key takeaways from this year’s Forecasting Benchmarking Study include:
Network complexity continues to outpace sales. Since 2010, the number of items for sale grew by 32% compared to a 4% increase in sales. As a result, average sales per item dropped 22%.
The rate of new product introductions is considerably higher, with the number of distinct items for sale nearly tripling in the last 5 years, up 187%; 82% of these have since been discontinued. Only 5% of all new items generate sufficient sales in their first year of production to escape being relegated to the tail.
The long tail continues to grow longer and now accounts for 81% of items. A different view of the data shows that the top 10% of items generate 75% of sales; whereas the bottom 50% only contribute 1%.
Existing demand planning technology and processes have reached their limits, with forecast value- added and accuracy remaining essentially flat, varying by no more than +/- 2% since 2010. This is a tribute to planners who have managed to maintain the status quo despite rapid proliferation, but is not sustainable.
Demand Sensing provides a step-change in performance that management seeks by automating the use of market data, more than doubling forecast value-added and cutting average forecast error by 37%.
The data confirms Demand Sensing enables a step-change in performance, more than doubling forecast value-added.