Forecasts are never right but you can improve your forecast accuracy …

1. You can’t improve what you don’t measure
2. Aggregate – individual plus or minus swings cancel each other out
3. Measure forecast accuracy at the right time fences
4. Review forecasts by exception only.  If the forecast deviation is within expected limits leave well enough alone
5. Tolerance – decide what ‘good’ and ‘bad’ forecast means
6. Triangulate – use multiple forecast methods
7. Automate and remove the human factor
8. Forecast less – less often, less granular
9. Demand Forecast Segmentation – Forecast-ability vs. Benefit
10. Reduce supply lead time, and you won’t need to forecast (as much)