The verdict in three sentences
Payment fraud silently eats around 1.8% of GMV in African e-commerce — 18,000 FCFA lost per million collected. Three levers make the difference: velocity rules, 3D Secure (which cuts card fraud by ~60% at the cost of +8% friction), and device fingerprinting. A simple, transparent scoring engine is enough to block most fraud without choking conversion.
The signals that give fraud away
A fraudulent payment almost always leaves traces: the same card tested across several accounts, repeated round amounts, an IP inconsistent with the card's country, a burst of attempts in minutes. The scoring engine's job is to add up these signals and decide: accept, review, or block.
| Signal | Score weight | Action past threshold |
|---|---|---|
| 3+ cards on one account / 24h | High | Manual review |
| 5+ failed attempts / 10 min | High | Temporary block |
| IP country ≠ card country | Medium | 3DS challenge |
| New account + large basket | Medium | Manual review |
| Disposable email | Low | Watchlist |
| Device tied to past dispute | High | Block |
3D Secure, velocity and fingerprinting
3D Secure (OTP or bank-app validation) shifts fraud liability to the issuing bank and slashes card fraud. Downside: it adds a step and therefore abandonment. So you enable it smartly, mainly on risky baskets.
| Control | 2026 fraud effect | Friction effect | When to enable |
|---|---|---|---|
| Blanket 3D Secure | -60% card fraud | +8% abandonment | Baskets > threshold |
| Velocity rules | -30 to -45% card testing | near zero | Always |
| Device fingerprint | -25% multi-account | zero | Always |
| Greylist / block | -20% repeat abuse | zero | Always |
| Per-device daily cap | -15% abuse | low | Per activity |
*2026 orders of magnitude; effect depends on your payment mix.*
Mini case study
Elom runs an online electronics shop in Johannesburg, GMV 6M FCFA/month. At 1.8% fraud he lost 108,000 FCFA/month in unpaid orders and disputes. He enables 3D Secure above 150,000 FCFA, velocity rules and fingerprinting. Fraud drops to 0.6% of GMV, i.e. 36,000 FCFA — a saving of 72,000 FCFA/month. Targeted 3DS costs him only ~1.5% abandonment on large baskets, easily offset.
Become a Kolonell referral partner
A poorly protected merchant loses a salary's worth to fraud every month. If you know some, the Kolonell referral program pays you for the introduction: 15% + 5% recurring on showcase sites, 12% on e-commerce, 10% on marketplaces, 8% on institutional. One referral, one commission — recurring on ongoing projects.
Need a professional website?
Kolonell builds websites that attract clients, optimized for the Sénégalese market. Free quote in 2 minutes.
FAQ
How much does African e-commerce fraud cost?
Around 1.8% of GMV in 2026, i.e. 18,000 FCFA per million collected. A good setup often brings it below 0.7%.
Is 3D Secure worth the friction?
Yes, when targeted: it cuts card fraud by about 60% for +8% abandonment. Enabled only above a threshold, the net gain is strongly positive.
What is a velocity rule?
A cap on frequency: for example 5 failed attempts in 10 minutes triggers a temporary block. Near-zero friction cost.
Does device fingerprinting respect privacy?
Yes if it stays an anonymous technical identifier. It cuts multi-account abuse by about 25% without storing sensitive personal data.
Should everything be blocked automatically?
No: keep a manual review queue for mid-range scores so you don't reject real customers (10-15% of suspicious cases are legitimate).
Let's talk about your project. We'll install your anti-fraud engine and rules in a week. WhatsApp +221 77 596 93 33.
Mohamed Bah
Fondateur, Kolonell
Passionate about digital and entrepreneurship in Africa, Mohamed has been helping Sénégalese businesses with their digital transformation since 2020. Founder of Kolonell, he believes every SME deserves a professional and accessible online présence.
