The MAPE formula
For each period, take the absolute gap between forecast and actual, divide by the actual, then average across periods.
``` MAPE = average of ( |Actual - Forecast| / |Actual| ) x 100 ```
Worked across four quarters:
| Quarter | Forecast | Actual | Absolute % error |
|---|---|---|---|
| Q1 | $4.2M | $4.5M | 6.7% |
| Q2 | $5.1M | $4.8M | 6.3% |
| Q3 | $5.4M | $5.9M | 8.5% |
| Q4 | $6.8M | $6.2M | 9.7% |
| MAPE | 7.8% |
Where MAPE breaks down
MAPE divides by the actual, so it inflates whenever the denominator is small. Roll MAPE up from rep level to team level and the smallest territories drag the average even when the dollars at stake are trivial.
Two rules keep the number honest. Compute MAPE on the aggregate you actually act on rather than averaging rep-level MAPEs. And never compute it on a period where actuals can hit zero, because the formula has no defined answer there.
What MAPE hides
MAPE cannot see systematic tilt. A forecast that lands 12% high one quarter and 12% low the next posts the same MAPE as a forecast that lands 12% high every quarter. The first is noise. The second is a correctable defect. Pair MAPE with a signed bias measure so you know which one you are holding, and track both against your forecast accuracy baseline.
Setting a MAPE target
There is no universal bar. ORM sees manual forecasts on new and expansion business land near 90% accuracy, roughly 10% error, produced at a heavy cost in analyst time and static once built. ORM targets 95% accuracy, or 5% error, and holds it from day one to day 90 of the quarter without manual adjustment.
Build a baseline from four to eight quarters of history before setting a goal. A team measuring MAPE for the first time usually finds the number worse than expected, which is the right starting point for a sales forecasting rebuild. For the mechanics of producing the forecast the score grades, see how to forecast revenue.
Frequently Asked Questions
How do you calculate MAPE for a sales forecast?
For each period take the absolute gap between forecast and actual, divide by the actual, then average those percentages across periods. Four quarters with errors of 6.7%, 6.3%, 8.5% and 9.7% produce a MAPE of 7.8%.
What is a good MAPE for revenue forecasting?
The useful bar is the error level at which finance can commit spend without hedging. ORM sees manual forecasts on new and expansion business land near 90% accuracy, about 10% error, and targets 95% accuracy, or 5% error, with a model that updates through the quarter.
Why does MAPE look worse at the rep level than at the team level?
MAPE divides by the actual, so small territories inflate. A $30K miss against a $150K rep quarter scores 20% while the same $30K miss against a $3M team number scores 1%. Compute MAPE on the aggregate you actually act on rather than averaging rep-level scores.
What is the difference between MAPE and forecast bias?
MAPE uses absolute values, so it measures the size of the miss and ignores direction. Bias keeps the sign and tells you whether the forecast runs consistently high or low. A forecast can post a respectable MAPE while carrying a standing bias you should have corrected quarters ago.
Put these metrics to work
ORM builds custom revenue forecast models that turn concepts like mean absolute percentage error (mape) into prescriptive action for your team.
Schedule a Demo