A delivery fleet is a capacity bet. Own too few riders and you miss promises at peak; own too many and you pay people to sit outside restaurants. The right model changes with density, order mix and labour rules.
Fleet management is not just hiring. It is shift design, utilisation, maintenance, insurance, retention and a cost-per-delivery number you can defend.
| Model | Cost shape | Control | Best for |
|---|---|---|---|
| Owned fleet | Fixed salaries, vehicles, insurance | High | Dense zones, brand-critical service |
| Gig / per-delivery | Variable per delivery | Medium | Variable demand, early launch |
| Third-party fleet | Contract or per-delivery | Low–medium | New cities, overflow capacity |
| Hybrid | Fixed core + variable overflow | High | Most growing platforms |
| Vendor self-delivery | Restaurant bears cost | Low | Small vendors, pilot zones |
Start with utilisation, not headcount
Deliveries per rider-hour is the metric that decides fleet cost. If a rider completes 2.5 deliveries an hour, the cost per delivery is the hourly cost divided by 2.5. If batching or density raises it to 4, the cost per delivery falls without a pay cut.
Before adding riders, check whether the existing fleet is waiting at restaurants or idle between orders. Waiting time is often a prep-time or dispatch problem, not a capacity problem.
Shift design for meal peaks
Demand is not flat. Lunch and dinner peaks need more riders than the midday trough. Build shifts around order curves, not convenience. A split shift (lunch + dinner) can be more efficient than a long shift with two idle hours, but it depends on local labour rules and rider preferences.
Keep a small overflow pool for rain, promotions and sporting events. Paying a premium for overflow is cheaper than missing promises every Friday.
Owned vehicles versus rider-owned
Owned bikes or scooters give control and branding, but they add maintenance, insurance, parking and depreciation. Rider-owned vehicles shift those costs but make quality and safety harder to standardise. A mixed model is common: company vehicles for full-time riders, allowances for gig riders.
Whichever you choose, define who pays for fuel, repairs, fines and accidents. Ambiguity here causes disputes and downtime.
Maintenance and safety
- Pre-shift vehicle checks: brakes, lights, tyres, phone mount.
- Service schedule by distance, not calendar alone.
- Helmet and reflective gear as a condition of work.
- Accident reporting and insurance claim process.
- Spare batteries or fuel policy for long shifts.
- Weather policy: when to pause or reduce zones.
Rider retention
Rider churn is a fleet cost. Recruiting and training a new rider takes time, and inexperienced riders are slower and more error-prone. Retention improves when pay is predictable, shifts are respected, support answers quickly and safety is taken seriously.
Small things matter: a working app, clear addresses, a human on the phone when a customer refuses an order, and payment on the promised day. A rider who trusts the platform stays through a bad weather week.
Cost per delivery model
Add rider pay, fuel/vehicle, insurance, maintenance, phone/data, support and management time. Divide by deliveries in the period. Compare that number with the delivery fee and the commission contribution. If the delivery fee does not cover the cost, the order is being subsidised by the food margin — which may be fine, but it must be deliberate.
When to outsource
Outsource when entering a new city, when demand is too volatile to justify fixed shifts, or when a third party can supply vetted riders faster than you can hire. Keep a small core fleet for quality and use partners for overflow. Do not outsource the customer promise without setting service levels and penalties.
See the management software, the dispatch guide, or peak-season planning.