Most “startup cost” articles quote a custom app build and stop there. The app is rarely the largest cost. Rider acquisition, restaurant discounts, refunds, cash reconciliation and marketing usually decide whether the first six months work. The ranges below are planning estimates, not quotes.
| Market | Software | Server / month | Launch budget* | Main local cost |
|---|---|---|---|---|
| Bangladesh | $299 once | $10–20 | $1,500–$5,000 | Rider cash, COD reconciliation |
| India | $299 once | $10–20 | $2,000–$7,000 | Discounts, GST, rider supply |
| Indonesia | $299 once | $10–20 | $2,000–$6,000 | Wallet/QRIS integration, rider fuel |
| Egypt | $299 once | $10–20 | $1,500–$5,000 | COD float, address quality, Ramadan peaks |
| Nigeria | $299 once | $10–20 | $1,500–$5,000 | Fuel, cash, payment reliability |
| Germany | $299 once | €10–20 | €5,000–€20,000+ | Rider labour, compliance, German copy |
*Excludes the software licence and server. Ranges cover branding, first restaurant incentives, rider onboarding, launch marketing, legal/accounting and a small operating buffer. They vary widely by city and scope.
The costs people forget
- Payment gateway fees — per transaction, plus chargebacks and refunds.
- SMS and OTP — small per message, large at volume; failed OTPs look like a broken app.
- Maps and routing — free tiers run out quickly once riders are live.
- Rider cash float — cash-on-delivery orders tie up working capital.
- App store accounts — Google Play and Apple developer fees, plus review delays.
- Customer refunds — budget 1–3% of order value for the first months.
- Support — WhatsApp/phone coverage during meal peaks is a real payroll line.
How to cut costs without breaking the model
Launch one dense zone instead of a city. Sign 25–40 restaurants before spending on ads. Use vendor self-delivery where rider density is weak. Start with cash and one card gateway; add wallets when the volume justifies the integration. Buy software you can own instead of stacking monthly SaaS fees before the first order.
Read the local guides for Bangladesh, India, Indonesia, Egypt, Nigeria and Germany.
The full launch budget
Most startup-cost articles stop at the app. The app is rarely the largest line. A realistic launch budget has six buckets: software, infrastructure, supply acquisition, rider operations, marketing and working capital. The last two are the ones that decide whether the business survives the first six months.
| Bucket | What it includes | Typical share of launch budget |
|---|---|---|
| Software | Licence, installation, branding, app publishing | 5–15% |
| Infrastructure | Server, domain, SSL, SMS, maps, email | 3–8% |
| Supply | Restaurant recruitment, menu photography, launch incentives | 10–20% |
| Riders | Recruitment, training, equipment, first shifts, cash float | 20–35% |
| Marketing | Hyperlocal ads, creators, referrals, launch offers | 20–30% |
| Working capital | Refunds, chargebacks, cash-on-delivery float, buffer | 10–20% |
The shares shift by market. In Germany, rider labour and compliance dominate. In Bangladesh or Egypt, cash float and restaurant incentives may be larger. In Indonesia, wallet and QRIS integration and rider fuel matter. The point is to budget all six buckets, not just the licence.
Country-by-country assumptions
The table in the main guide gives planning ranges. Behind those ranges are different operating models:
- Bangladesh: cash-on-delivery is common, restaurant commissions are under pressure, and rider supply is dense in Dhaka but thinner in secondary cities. Budget for cash reconciliation and a launch zone, not a city-wide map.
- India: UPI is default, discounts are aggressive, and GST and FSSAI compliance add administrative cost. A niche or tier-2 launch is usually more capital-efficient than a metro discount war.
- Indonesia: GoFood and GrabFood dominate discovery; QRIS, wallets and cash matter. Outside Jakarta, rider supply and address quality are the constraints.
- Egypt: cash and Fawry/Paymob are part of the mix; address quality and Ramadan peaks shape operations. A dense Cairo cluster is the usual starting point.
- Nigeria: fuel, cash and payment reliability affect delivery cost. Dense Lagos zones can work; long-distance delivery rarely does.
- Germany: labour rules, insurance and compliance raise the fixed cost. A German launch needs a higher basket or a subscription model, not a $2 delivery fee.
The first six months
Month one is not profitable. You are buying density, learning the menu and fixing operational errors. A realistic plan keeps three months of rider and support costs in reserve, plus a refund buffer. If the first zone does not reach contribution break-even by month three, do not expand; fix the zone.
Track cash weekly, not monthly. Delivery businesses can look profitable on an accrual report while running out of cash because rider payouts, refunds and restaurant settlements land on different days. A simple 13-week cash forecast is more useful than a detailed annual budget.
Funding and cash flow
Bootstrapped operators should avoid fixed rider salaries until order density is proven. Use per-delivery pay or vendor self-delivery for the pilot, then convert the best riders to shifts. If you raise money, do not use it to subsidise every order; use it to buy density, technology and restaurant supply in one zone.
Prepaid wallets and subscriptions can improve cash flow, but they create a liability. Track unredeemed balances and do not spend them as revenue. The same applies to restaurant security deposits and rider cash floats.
Common budgeting mistakes
- Budgeting the licence but not the app-store accounts, SMS, maps and payment fees.
- Launching five zones because the map looks empty.
- Hiring riders before restaurants and demand are ready.
- Discounting every order instead of funding a first-order offer.
- Ignoring refunds, failed deliveries and chargebacks.
- Treating prepaid wallet cash as profit.
Fix those six and the launch budget becomes a plan rather than a guess.