A forecast of '1,200 units' sounds precise and is almost always wrong by some amount. Decision-makers who see only the point estimate plan as if it were certain.
Practices
- Show a range with a stated confidence level, and explain it in plain words.
- Give two or three named scenarios, with what would have to be true for each.
- Track forecast accuracy over time and publish it, so trust rests on evidence.
- Tie ranges to actions — what you do at the low end, and at the high end.
A forecast is useful when it changes a decision. If the same action is right across the whole range, the precision doesn't matter.