Skip to content

Food-delivery riders in a metro

Workforce sizing from order volume and rider throughput.

moderate
5 min read
workforcedemand-capacity

Estimate the number of active food-delivery riders in a 10-million metro. Workforce guesstimates are demand ÷ throughput problems with a peak-hour correction — fleets are sized for dinner rush, not the daily average.

Users → orders → the shape of the day (size for the dinner spike, not the average) → throughput → roster gross-up.
1

Orders

10M people → ~2.5M food-delivery users (urban, smartphone, ordering habit); ~0.25 orders/user/day blended → ~600K orders/day.

2

Peak concentration

Dinner 8–9 pm carries ~18% of daily volume → ~110K orders in the peak hour.

3

Throughput

At peak with order-stacking, a rider completes ~2.5 deliveries/hour → need ~45,000 riders on-shift at peak.

4

Active base

Part-time mix and shift patterns mean the weekly-active base ≈ 1.6× peak requirement → ~70–80K riders.

600k × 18% ÷ 2.5/hr ≈ 43k at peak × 1.6 roster factor ≈ ~70k active riders

How to defend it

The peak-sizing step is the whole point — sizing on daily averages gives ~25K and would mean hour-long dinner waits. Any workforce/fleet guesstimate (cabs, ambulances, servers) should be sized at peak and reconciled to roster.