2011 Forecasting Performance Benchmark Study
Naïve Forecast Error
The naïve forecast has been upgraded to a simple seasonally-adjusted moving average forecast to create a more meaningful comparison with Demand Planning. Though less naïve, it still lacks Demand Planning’s ability to reflect promotions or expert knowledge from planners. Furthermore, it lacks the myriad of sophisticated statistical time-series analysis options available in Demand Planning software solutions.
With the high number of promotions in 2010, Demand Planning should well outperform the simplistic naïve forecast. The value of planners is evident at longer horizons, with Demand Planning beating the simplistic model by 22% for monthly forecasts, and as high as 36% for Horizon 56. However, the effect diminishes as the horizon shortens, with Demand Planning and naïve forecast error converging at Horizon 7.
This convergence suggests that planner’s time is better spent focused on longer range activities where the value of Demand Planning is greatest and that short-term forecasts are better left to Demand Sensing which consistently outperforms both Demand Planning and the naïve forecasts.
