The verdict in three sentences
Merchant-side mobile money fraud does not require sophisticated AI: a few simple rules block 70 to 85 % of attempts. In 2026, fake payment receipts account for about 3 % of orders and abusive refunds 1.5 %, two leaks most merchants pay without seeing. The golden rule: never ship on a screenshot, only on a signed-webhook confirmation.
The fraud signals to watch
Most fraud follows repetitive patterns. A lightweight rules engine wired to your orders assigns a risk score and holds anything above a threshold. Here are the most cost-effective signals to watch.
| 2026 signal | What it catches | Recommended action |
|---|---|---|
| Fake receipt (screenshot) | payment never received | require signed webhook |
| High velocity | same number/card in bursts | throttle and verify |
| Name / number mismatch | account takeover | ask for confirmation |
| Repeated round amounts | fraud test / laundering | manual flag |
| New address + big basket | mule / resale | enhanced verification |
| Immediate post-delivery refund | abusive refund | window + return proof |
Impact of rules on fraud
Every rule, even basic, removes a share of risk. The combined effect of a small bundle of rules beats any single rule.
| Rule applied | Estimated fraud reduction | Setup effort |
|---|---|---|
| Signed-webhook confirmation only | 40 to 55 % | low |
| Velocity limit per number/IP | 10 to 15 % | low |
| Name/number mismatch check | 8 to 12 % | medium |
| Refund window + proof | 10 to 15 % | low |
| Combined risk score | 70 to 85 % cumulative | medium |
The signed webhook alone removes the most common fraud in West Africa: the fake payment screenshot. Until your system receives the operator's signed notification, the order stays "unpaid".
Mini case study
Cheikh sells sneakers in Dakar, 500 orders/month at a 25,000 FCFA average basket (12,500,000 FCFA). He suffered ~ 3 % fake payments (15 orders, 375,000 FCFA shipped for nothing) and ~ 1.5 % abusive refunds (~ 190,000 FCFA), i.e. ~ 565,000 FCFA/month in losses. After enforcing signed-webhook confirmation and a refund window with return proof, he cuts fraud by 80 %, bringing losses to ~ 113,000 FCFA/month: ~ 452,000 FCFA recovered every month for a rules module at around 700,000 FCFA, paid back in under two months.
FAQ
Need a professional website?
Kolonell builds websites that attract clients, optimized for the Sénégalese market. Free quote in 2 minutes.
Isn't a payment screenshot enough proof?
No. Screenshots are faked in seconds. The only reliable proof is the signed-webhook confirmation received directly from the operator, which you cannot forge.
Will velocity checks block real customers?
Rarely, if thresholds are well calibrated. You hold for verification, not outright reject; a real customer confirms in one message and goes through.
How do I handle abusive refunds?
Set a refund window (say 48 to 72 h) and require proof of return or non-conformity. The dispute reserve covers the case where you already paid the vendor.
Do these rules slow down ordering?
No for clean orders, which pass in real time. Only risky orders, a minority, go to a few-minute manual review.
Do I need a big budget to protect myself?
No. A simple rules engine (2026 ballpark 500,000 to 900,000 FCFA) is enough to cut most fraud; advanced AI only becomes useful at very high volume.
Let's talk about your project. We audit your orders and switch on the rules that cut the most fraud on your store. 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.