Commission is the price of demand. Every restaurant compares it to the cost of acquiring that order themselves — a Facebook ad, a flyer, a discount — and decides whether you are worth it. Get the model wrong and you either subsidise every order or lose your best vendors.
Flat percentage
The default: 15–30% of the food subtotal, delivery fee excluded. It scales with basket size, which feels fair to both sides, and it is easy to explain. The weakness is that it punishes high-volume, low-margin vendors — a biryani house doing $6 plates feels 25% much more than a sushi bar doing $40 plates.
Use it as the base rate, then adjust by category rather than by favourite vendor.
Per-order fee
A flat amount per order, say $0.50–$1.50. Predictable for the vendor, and it protects you on small baskets where percentage commission does not cover dispatch. The failure mode is that your revenue stops growing with order value, so a platform with rising baskets ends up undercharging.
Tiered percentage
Lower rates for high volume, or for vendors who handle their own delivery. This is the most common real-world structure and the easiest to defend in a negotiation: "you keep more per order as you grow with us". Keep the tiers few — three is plenty — and publish them.
Subscription
The vendor pays a fixed monthly fee for unlimited orders. Attractive to busy restaurants, dangerous for you: you carry dispatch cost on every order while your revenue is capped. It works only when the vendor is self-delivering, or when the fee is priced against their actual order volume.
Hybrid
A lower percentage plus a per-order fee, or a percentage plus a delivery fee retained by the platform. This is where most mature platforms land, because it separates the two things you are selling: demand (percentage) and logistics (fee).
What to do with the delivery fee
Decide early who keeps it. If the customer pays $2 delivery and the platform keeps it, your commission can look lower and vendors accept it more easily. If the vendor keeps it, your commission has to carry the fleet. Whichever you choose, write it down — this is the single most common source of settlement disputes.
The retention test
A useful sanity check: would your top ten vendors stay if a competitor offered 5 points less? If the answer is no, you are renting supply with price, and the only exit is demand density — being the channel that actually brings customers. That is a marketing problem, not a commission problem.
Changing commission later
Announce at least 30 days ahead, grandfather existing vendors for one cycle, and tie the increase to something visible: more riders in their zone, a new customer-acquisition campaign, faster payouts. A surprise increase reads as a cash grab and costs you the vendors who care most about margin — usually the reliable ones.
The maths behind a 15% commission
A 15% commission on a $25 order is $3.75. If the rider costs $2.50, the payment fee is $0.63 and packaging/support is $0.50, the contribution is $0.12 before fixed costs. That is not a business. The commission rate cannot be evaluated in isolation; it only works with the basket size, delivery fee and rider cost.
| Order value | 15% commission | After $2.50 rider + 2.5% fee |
|---|---|---|
| $15 | $2.25 | −$0.63 |
| $25 | $3.75 | $0.62 |
| $35 | $5.25 | $1.87 |
| $50 | $7.50 | $3.50 |
The same 15% is unprofitable on a $15 order and healthy on a $35 order. That is why minimum baskets, combos and delivery fees exist. It is also why a commission model that ignores basket size is incomplete.
What restaurants compare
Restaurants do not compare only the percentage. They compare the payout they actually receive, the settlement date, who pays for delivery, who handles refunds and how much marketing the platform brings. A 20% commission with reliable weekly payouts and real orders can be better than a 12% commission with late settlements and no demand.
When you change commission, show the restaurant the old and new payout on a real order. A percentage change sounds abstract; a lower payout on last week's orders is concrete. Give notice, explain the reason and offer a path to a lower rate through volume or self-delivery.
How to change commission without churn
- Segment vendors: high-volume, growth, low-volume and self-delivery.
- Model the new payout per vendor, not just the platform revenue.
- Give at least 30 days' notice and a clear effective date.
- Offer a retention path: volume tiers, self-delivery discount or a marketing package.
- Track vendor churn and order volume for 90 days after the change.
A commission calculator
Build a simple spreadsheet: order value, commission %, delivery fee, rider cost, payment fee %, refund rate, packaging cost and support cost. Calculate contribution per order at $15, $25, $35 and $50 baskets. Then change the commission rate until the contribution is positive at your median basket, not your best one. If it only works at a $50 basket, your model is targeting the wrong customer.