A blunt guide to online payments in Africa

Africa has 54 countries. That matters. A checkout flow that works in Nairobi will fail in Lagos. Cards win in some markets. Bank transfers or mobile money win in others. Meet people where they already pay. Period.

Here's what I wish someone told me on day one.

What kills checkout conversion (real examples)

• Startup launches card-only in Kenya. 70% cart abandonment. Adds M-Pesa. Conversion jumps 40%.

• SaaS company routes Nigeria through European acquirer. 2-week settlement delays. Churn spikes.

• E-commerce site assumes one PSP covers "Africa." Goes live in 3 countries. Works in 1.

• Gaming app builds for smartphones only. Discovers 60% of users still on feature phones with USSD.

• Marketplace hard-codes USD in database. Launches in 4 countries. Reconciliation breaks immediately.

Sound familiar? Good. Now let's fix it.

The basics that decide your success

Rails you'll meet over and over

• Mobile money: wallets run by telcos (M-Pesa, MTN MoMo, Airtel Money, Orange Money). M-Pesa processes more transactions than Visa and Mastercard combined in Kenya. GSMA shows Sub-Saharan Africa drives global mobile money growth.

• Account-to-account (A2A) / instant bank transfers: Nigeria's NIP, South Africa's PayShap/RTC, Egypt's InstaPay. Nigeria's NIP processes 50x more transactions than all card schemes combined.

• Cards: stronger in South Africa, Egypt, Morocco, and for cross-border shoppers almost everywhere. You still need Visa/Mastercard in any pan-African stack.

Why cross-border breaks everything

• Fragmented standards: many IPSs still on ISO 8583; newer ones moving to ISO 20022. ISO 8583↔20022 translation will break somewhere in your stack. Plan for it.

• Regional schemes:

  • PAPSS aims to clear+settle cross-border in local currencies. Coverage expanding fast but adoption remains uneven and stepwise.

  • BUNA serves Arab-African countries with real-time cross-border in multiple currencies (USD, EUR, AED, EGP, SAR, JOD).

• FX and compliance: capital controls, documentary requirements, and grey-listing slow or reroute flows. FATF monitoring affects multiple African markets. Expect bank-level scrutiny on these corridors.

What customers actually use (reality check by market)

Kenya

• What works: M-Pesa STK push → PesaLink A2A → Cards

• Reality: M-Pesa owns 90%+ of mobile money lines. Lipa Na M-Pesa merchant base keeps growing. Design for STK flows or lose customers.

Nigeria

• What works: NIP bank transfer → Cards/USSD

• Reality: NIP is the dominant digital rail. Volumes and values at record highs. If you can't take bank transfers, you don't really sell online in Nigeria.

South Africa

• What works: Cards + Instant EFT → PayShap

• Reality: PayShap launched 2023 with small-ticket limits (ZAR 3,000 per tx). RTC still runs on ISO 8583, PayShap on ISO 20022. QR apps (SnapScan, Zapper) add POS coverage.

Egypt

• What works: Cards (Meeza) → InstaPay A2A → Wallets

• Reality: InstaPay exploded in usage. Meeza cards widespread. BUNA offers real-time cross-border for eligible currencies.

Ghana

• What works: Mobile Money (MMI interop) → Cards

• Reality: GhIPSS MMI volumes rising. Interop is mature. Mobile money accounts keep climbing.

Uganda

• What works: Mobile Money (MTN/Airtel) → Cards

• Reality: Two players dominate mobile money lines. USSD+STK essential. Cards are niche.

Morocco

• What works: Cards → Instant credit transfer (Virement Instantané)

• Reality: New ISO 20022 instant rail live since 2023 with limits.

What you'll actually spend

Multi-PSP integration: 3-6 months engineering time

Per-country compliance: $10K-50K legal/regulatory setup

Cross-border FX spread: 2-5% hidden in "competitive" rates

Reconciliation headaches: 0.5-1 FTE ongoing

Payment method testing: Budget for cheap Android phones and 2G networks Webhook infrastructure: Replay systems aren't optional

Infrastructure that will break (and when)

USSD timeouts during peak hours: Payday weekends kill USSD flows

Webhook failures during telco maintenance: Happens every weekend somewhere

Currency conversion rounding errors: Compound across high-volume flows

Bank API rate limits: Kick in without warning during traffic spikes

Reference number mismatches: Different PSPs use different reference formats Delayed settlement notifications: Can arrive hours or days after transaction

The stack that gives you widest reach

Red flags when choosing PSPs

If they can't show you test credentials in 24 hours, walk away. If settlement reports don't match transaction logs, you'll hate reconciliation. If they quote "all of Africa" coverage, they're lying or naive. If webhook delivery isn't idempotent, your reconciliation will break. If they can't explain their FX markup clearly, it's too high.

1) Elemipay (by Elemitech) — aggregator built for African fragmentation

• What it is: API-first acceptance across mobile money, bank transfers, and cards. One integration for local rails.

• Best for: SaaS, marketplaces, fintech apps needing mobile money + bank rails across East/West Africa without stitching many PSPs.

2) DusuPay — mobile money + cards across regions

• Strengths: Strong MoMo coverage (M-Pesa, MTN, Airtel, Orange, etc.), plus cards. Collections and disbursements.

• Best for: Apps that live on MoMo (rides, fintech, digital goods, remittances).

3) Flutterwave — pan-African acquiring + payouts

• Strengths: Broad geographic footprint. Cards, bank transfers, mobile money. Robust payouts. Strong enterprise references.

• Best for: Scale-ups and enterprises needing breadth and treasury tools.

4) Paystack — developer-loved, strong in NG/GA/ZA/KE

• Strengths: Clean APIs and docs. Great checkout UX. Cards + bank transfers + EFT. Solid risk tools.

• Best for: Startups to mid-market merchants wanting speed to market in these four markets.

5) Onafriq (ex-MFS Africa) — the "network of networks"

• Strengths: Massive mobile money and account connectivity. 40+ markets, 500M+ wallets and 200M+ bank accounts reachable.

• Best for: Platforms that must pay in/out across many countries more than classic card acquiring.

How to choose fast

• Need Kenya, Tanzania, Rwanda, Uganda first? Make mobile money your default.

• Need Nigeria first? Make NIP bank transfers your default.

• Need South Africa? Lead with cards + instant EFT.

• Need Egypt? Lead with cards (Meeza) and InstaPay A2A.

FULL GUIDE TO WHICH PSP YOU CAN CHOOSE: Top Payment Gateways in Africa: The Ultimate Guide

Architecture that actually works

1) Go multi-PSP, single-checkout. Abstract providers behind one UX. Route by country/rail. Fail over. No single PSP gives you Africa-wide reliability. None.

2) Build per-country defaults.

• Kenya/Uganda/Tanzania/Rwanda → mobile money first

• Nigeria → NIP first

• South Africa → cards/instant EFT first

• Egypt → cards + InstaPay

This alone lifts checkout conversion 10-30% vs card-only flows.

3) Design for brittle standards. Budget for message mapping, idempotency, and reconciliation quirks.

4) Decide cross-border strategy early.

• Low corridor volume? Use PSP payouts.

• Strategic corridor? Explore PAPSS via your bank/PSP once live. Expect staged onboarding.

5) Treat KYC/KYB as a product. Grey-list status, FX paperwork, and tax rules differ by country. Bake document capture into onboarding. Keep settlement currencies flexible.

Mistakes that will cost you months

• Launching without USSD fallbacks in mobile money markets

• Hard-coding currency in your database schema

• Treating failed payments as declined cards

• Building country detection from IP (VPNs will break this)

• Assuming webhook order matches transaction order

• Not testing on 2G networks with cheap phones

• Skipping payment method fees in product pricing

• Building reconciliation as an afterthought

What experienced teams do differently

• Test payment flows on 2G networks with cheap Android phones

• run parallel PSPs for 30 days before switching primary routing

• We price payment method fees into product pricing, not checkout

• We built internal tooling to replay failed webhook deliveries

• We soft-fail to local alternatives when primary rail is down

• We reconcile daily, not monthly

• We explain fees and timing in plain language during checkout"

Implementation notes that save quarters

• USSD and offline flows aren't optional. Don't assume 4G or smartphones. 60% of users in some markets still use feature phones. Design "pay by USSD" entry points.

• Soft-fail, then educate. Rail down? Auto-offer local alternative in-flow. Explain fees and timing upfront.

• Reconcile like an adult. Expect reference mismatches, delayed webhooks, currency rounding. Build a reconciliation service, not a cronjob.

• Price intelligently. Fees swing from 0.5% to 3%+ by rail and country. Don't hard-code take rates. Pass-through where it makes sense. Subsidize when conversion is king.

• Plan for upgrades. More IPSs migrating to ISO 20022. Keep message models clean.

---

GET STARTED ACCEPTING PAYMENTS FROM ALL OVER AFRICA

Learn more

---

Final word

Africa is not one market. Treat each country like a product. Lead with the local rail. Add cards for cross-border. Use at least two PSPs behind one checkout.

Pick a partner that speaks mobile money and A2A, not just cards.

If you want, I can tailor this to your exact go-live list and draft the integration map (per country, per rail, per PSP) in one pass.