S&OP, Demand Planning & Forecasting worked example
Forecast Bias at 99% plan attainment: a worked example
What does the result look like when plan attainment reaches 99%? The full calculation is worked below with real intermediate numbers. Use it when forecast bias in s and op, demand planning and forecasting is being asked to take on more work and you need to know if there is room.
The inputs for this scenario
- Net forecast error per planning cycle: 4 units / cycle (unchanged)
- Planning cycles in the review window: 480 cycles (unchanged)
- Plan attainment (uptime equivalent): 99 % (raised for this scenario; the documented default is 90)
- Forecast-to-actual pass rate: 97 % (unchanged)
Working through the calculation
- Applying the documented formula (Gross forecast bias capacity = forecast bias output per cycle × available forecast bias cycles) to the inputs above produces each figure below.
- At this operating point the engine returns 1,844 units for good forecast bias capacity, the number this scenario is built around.
- At this operating point the engine returns 1,920 units for gross forecast bias capacity.
- At this operating point the engine returns 19.2 units for forecast bias downtime loss.
- At this operating point the engine returns 57.02 units for forecast bias yield loss.
How this compares with the baseline
- Against the tool's baseline example, where plan attainment sits at 90% and the headline result is 1,676 units, this scenario comes in 10% above the baseline at 1,844 units.
- A figure at this level is achievable when plan attainment is genuinely sustained, not just peaked for a shift. It treats bias as a steady per-cycle quantity; if bias swings direction between cycles this overstates the net effect.
Results at a glance
- Good forecast bias capacity: 1,844 units (headline result)
- Gross forecast bias capacity: 1,920 units
- Forecast bias downtime loss: 19.2 units
- Forecast bias yield loss: 57.02 units
Run it with your numbers
- Every input above is editable in the live Forecast Bias calculator, which recalculates instantly and can be shared with the inputs intact.
Last reviewed 2026-05-12.