Last-mile delivery startup Africa: 2026 tech stack (Glovo-killer)

Glovo, Yango Food, Heetch Food dominate last-mile delivery in African capitals. But their commission (22-28%) chokes restaurants and shops. A local last-mile startup, focused on 1-2 cities, can be profitable from 200-500 orders/day with modern tech stack.

TL;DR

- Stack: Next.js apps (customer + driver + merchant) + Postgres + Mapbox + Wave.

- Glovo differentiation: 12-15% commission + better merchant UX + native Wave/OM payment.

- Market: Dakar, Abidjan, Douala, Lagos; secondary (Bamako, Cotonou, Ouaga).

3-app architecture

`

[Customer app: order + payment]

↓

[Merchant app: receives + prepares + ready]

↓

[Driver app: pickup + delivery]

↓

[Backend orchestration]

├── Smart routing (Mapbox Optimization API)

├── Dynamic pricing (zone + time)

├── Wave / OM / cash payment

└── WhatsApp + push notifications

`

Data model

`prisma

model Merchant {

id String @id @default(cuid())

type String

businessName String

address String

gpsLat Float

gpsLng Float

hours Json

preparationTimeAvgMin Int

commissionRate Float

walletProvider String

walletNumber String

isActive Boolean

}

model Driver {

id String @id @default(cuid())

firstName String

lastName String

phone String

whatsapp String?

vehicleType String

registrationNumber String?

cniScan String?

walletProvider String

walletNumber String

rating Float?

totalDeliveries Int @default(0)

isActive Boolean

currentGps Json?

currentStatus String

}

model Order {

id String @id @default(cuid())

customerId String

merchantId String

driverId String?

items Json

subtotal Int

deliveryFee Int

commission Int

total Int

status String

paymentMethod String

paidAt DateTime?

pickupGps Json

deliveryGps Json

estimatedDeliveryAt DateTime?

actualDeliveryAt DateTime?

customerRating Int?

driverRating Int?

}

`

Smart routing

`tsx

async function dispatchOrder(orderId: string) {

const order = await getOrder(orderId);

const merchant = await getMerchant(order.merchantId);

const candidates = await prisma.driver.findMany({

where: {

currentStatus: 'ONLINE',

isActive: true,

},

take: 10,

});

const ranked = candidates

.map(d => ({

driver: d,

score: computeDispatchScore(d, order),

}))

.sort((a, b) => b.score - a.score);

for (const { driver } of ranked.slice(0, 3)) {

const accepted = await offerOrderToDriver(driver, order);

if (accepted) {

await prisma.order.update({

where: { id: orderId },

data: { driverId: driver.id, status: 'ACCEPTED' },

});

return;

}

}

}

`

Dynamic pricing

`ts

function computeDeliveryFee(order: OrderInput) {

const baseRate = 1000;

const distanceKm = computeDistance(order.merchantGps, order.deliveryGps);

const distanceFee = distanceKm * 200;

const hour = new Date().getHours();

const isPeak = (hour >= 11 && hour <= 14) || (hour >= 18 && hour <= 22);

const surgeFactor = isPeak ? 1.4 : 1.0;

const isRaining = await checkWeather(order.merchantGps);

const weatherFactor = isRaining ? 1.2 : 1.0;

const total = (baseRate + distanceFee) * surgeFactor * weatherFactor;

return Math.round(total / 100) * 100;

}

`

Real case — Dakar startup (Gora Express)

Common pitfalls

• Compete on price vs UX — beating Glovo by 5% cheaper is hard. Beating on merchant UX + native Wave payment more defensible.
• Driver churn — instant wallet recharge (not T+7) = loyalty.
• No instant CS — every delivery incident needs a human. Invest.
• Critical density not reached — under 80 orders/day, drivers demotivated.
• Peak-hour app crashes — Locust load testing + Sentry observability critical.

FAQ

Q: Initial investment?

A: 80M-200M XOF for solid MVP + 12 ops months. Investors: 4DX Ventures, Norrsken22, Partech Africa.

Q: Competition?

A: Glovo, Yango, Heetch. But secondary markets (Bamako, Ouaga, Lomé) less saturated.

Conclusion

Africa last-mile delivery startup 2026 = profitable local niche but demanding clean execution. Direct Glovo battle = death. Secondary niche or merchant differentiation = viable.