Setting up your own delivery removes the 25–35% commission that aggregators take on every order. For a restaurant doing $20,000/month in delivery, that's $5,000–7,000 staying in your pocket instead of going to Glovo or Wolt.
The tradeoff: you manage the logistics. Drivers, routes, order timing, and customer communication become your responsibility instead of the platform's.
This guide covers what it takes to run in-house delivery well — and when it makes sense vs. staying on aggregators.
When in-house delivery makes sense
It makes sense when:
- You have an established customer base who orders repeatedly (not just occasional new customers)
- Your delivery area is compact (under 5km radius)
- You have enough volume to justify dedicated driver time (15+ orders/day minimum)
- You're in a category where customers will seek you out directly (specialty cuisine, premium market, loyal regulars)
It doesn't make sense when:
- You rely heavily on aggregators for new customer discovery
- Your order volume is low and unpredictable
- Your delivery area is spread out (large distances between customers)
- You have no system for driver management
Many restaurants run a hybrid: aggregators for discovery and new customers, own delivery for established customers who order directly. This is often the optimal model.
Step 1: Define your delivery zone
Start small. A 2–3km radius from your restaurant is manageable for one or two drivers and keeps delivery times predictable.
Factors to consider:
- Delivery time commitment: Can you deliver to the edge of the zone in under 30–40 minutes? If not, shrink the zone.
- Driver efficiency: A tighter zone means drivers can complete more deliveries per hour.
- Food quality: Delivery time directly affects food quality. Soups, fried items, and anything temperature-sensitive degrade after 30 minutes.
Draw your zone on a map. Be honest about what's realistic. You can expand later once the operation is running smoothly.
Step 2: Decide on driver employment model
Option 1: Employed part-time drivers
You hire 1–2 part-time drivers for your peak hours (lunch and dinner service). They're on a fixed schedule, use their own or a restaurant vehicle, and are paid hourly or per delivery.
Pros: Reliability, control over presentation, drivers represent your brand. Cons: Labor cost regardless of order volume. Payroll, insurance, HR.
Option 2: Contracted per-delivery drivers
You work with local freelance couriers who pick up orders on demand. Apps like Stuart, Shipday, or local courier networks handle dispatch.
Pros: No fixed labor cost — you pay only when there's an order. Flexible. Cons: Less control over delivery experience. Reliability varies.
Option 3: A regular driver as part of the kitchen team
For low-volume operations (10–15 deliveries/day), a kitchen staff member with a scooter or car can handle deliveries between their other duties during off-peak hours.
Pros: No dedicated hire. Familiar face. Cons: Can't scale. Creates conflicts during rush periods.
Most independent restaurants start with option 3 or a single contracted driver, then graduate to employed drivers as volume grows.
Step 3: Set up your ordering system
You need a system to receive orders, communicate with drivers, and track status. Manual phone ordering works at very low volume but breaks down fast.
What you need:
- A way for customers to place orders (website, WhatsApp, phone)
- A way to receive and confirm orders in the kitchen (tablet, POS, notification)
- A way to assign and track deliveries (driver app, phone call, WhatsApp group)
Tools for small operations:
Restmarket: Your restaurant website receives orders. The kitchen gets a real-time notification on a dashboard. Orders show customer address and details. You coordinate with your driver via WhatsApp or phone.
Shipday: A dedicated delivery dispatch tool. Drivers get a mobile app with order details and route. You can track driver location. Free plan for small volume.
WhatsApp groups: Many small restaurants use a WhatsApp group between kitchen and driver. Order comes in, gets posted to the group, driver picks up. Low tech, works surprisingly well for 15–20 orders/day.
Step 4: Define the customer experience
Order confirmation: Customer should receive confirmation immediately after placing an order. Include estimated delivery time.
Delivery time estimate: Be conservative. "30–45 minutes" is better than promising 25 and delivering at 40. Customers who receive earlier than expected are delighted; late deliveries generate complaints.
Tracking: Customers expect some visibility. At minimum, send an update when the order leaves your kitchen ("Your order is on its way"). Live GPS tracking is ideal if your tools support it.
At the door: Brief, professional interaction. A bag with your branding, a receipt, and ideally a card promoting direct ordering / loyalty program.
Issue handling: Have a clear policy for wrong orders, cold food, or delayed delivery. A partial refund or discount on the next order resolves most complaints immediately. Make it easy for drivers to communicate issues back to the kitchen.
Step 5: Packaging for delivery
Food packaged for pickup at the restaurant is not always right for delivery. Think through:
Temperature retention: Insulated bags for hot food. Cold items separate from hot. Soup in sealed containers, not cups with lids that leak.
Travel stability: Tall items tip. Layered containers shift. Test how your packaging survives a 15-minute scooter ride before going live.
No-leak verification: Liquid-heavy dishes (soups, curries, sauces) need sealed containers. A single leak complaint requires addressing the packaging immediately.
Your branding: Bags, stickers, or stamps with your logo and website URL. Every delivery is a marketing moment.
Pricing for your own delivery
When you remove the aggregator, you can choose how to price:
Option 1: Same price as app, no delivery fee You pass all savings to the customer as lower effective price. Maximum incentive to order directly.
Option 2: Same price as app, small delivery fee ($2–3) Customer pays slightly less than on the app (where apps add their own fees), you cover driver costs.
Option 3: Lower prices + delivery fee Most transparent model. Lower menu prices (you're not paying 30% to the app), modest delivery fee.
Most restaurants find option 2 or 3 works well. The key message to customers: "Order directly — you pay less, we earn more."
Calculate your delivery cost per order (driver wage + fuel + amortized packaging) before setting the delivery fee. A typical short-distance delivery costs $2–5 in driver time.
Minimum order value: Set a minimum that makes delivery economically viable. If your delivery cost is $4 and minimum order is $8, you're barely covering logistics. A minimum of $15–20 for delivery is reasonable for most restaurants.
Managing peaks and slow periods
Delivery volume is uneven. Friday evening is 3x Monday lunch.
Tactics:
- During rush, tighten your delivery zone (reduce to 1.5km to keep times consistent)
- Add a "high demand" estimated time warning during peak hours
- Pre-schedule extra driver coverage for Fridays, weekends, lunch peaks
- Consider pausing direct delivery during the most intense aggregator hours if you can't staff both
Measuring your own delivery operation
Track weekly:
- Orders per day by channel (own vs. apps)
- Average delivery time (from kitchen out to delivered)
- Issue rate (complaints, wrong orders, late deliveries)
- Cost per delivery (driver cost / number of deliveries)
- Revenue per delivery (gross revenue - packaging - driver cost)
Compare revenue per delivery on your own channel vs. aggregator channel. The aggregator number should be 25–35% lower (the commission you're not paying). That gap is what drives the case for expanding your own channel.