One quick staffing decision

There are no right answers. We're after your read of the situation.

You manage staffing for a retail showroom. A forecasting model estimates demand and recommends how many people to schedule.

Here is the model's recommendation over the last 15 days, next to the number of staff actually needed:

Model recommendation Actual need

Show the numbers

Would you use the model as the primary basis for tomorrow's staffing plan, or rely on your own judgment instead? Use the model as the primary basis Rely on my own judgment instead

How many staff would you actually schedule?

Why?

MAPE tolerance

Forecasters measure how good a model is with a number called MAPE — the model's average miss, written as a percentage. If the model says “schedule 10” and the day really needs 9 or 11, that’s a 10% miss. Do that across many days and the typical miss is the MAPE.

So a MAPE of 10% means the recommendation is normally about one person out of ten away from what the floor actually needed — sometimes over, sometimes under. A MAPE of 25% means it’s routinely off by a quarter.

The worst MAPE you’d still be willing to staff from (%) Above this number, you’d stop following the model and go with your own read instead. What makes that your cut-off?

Which best describes your retail experience? No retail experience Worked in retail but did not staff personally Was responsible for retail staffing

Optional (all three can be skipped)

Age (optional)

Gender (optional)

Country or region (optional)

Thanks, that's it.

Your answer was recorded. This is part of a study comparing how people and AI models decide when a forecast stops being worth using.

Everyone sees a slightly different version of the scenario. That variation is the whole point, and it's why the answer isn't obvious.

Consent to participate in research

One quick staffing decision

Optional (all three can be skipped)

Thanks, that's it.

You've already taken this one.