The verdict in three sentences
On MTN MoMo and M-Pesa, true card-style direct debit does not yet exist universally: you operate between tokens/mandates (when available) and push-payment with dunning. Your number-one risk is not the transaction fee (1.5 to 3%) but the involuntary churn of 8 to 15% per month caused by failed payments. A serious dunning engine recovers 30 to 40% of those failures and transforms your SaaS retention.
Token/mandate vs push-payment: the wallet matrix
Two models coexist. Tokenised auto-debit: the customer authorises once, you replay the charge monthly with no action from them. Push-payment: at each cycle the customer actively approves (USSD, notification, STK push). The latter is more resilient in African markets but generates abandonment.
| Rail (2026 order of magnitude) | Tokenised debit/mandate | Push + dunning | Fee per charge |
|---|---|---|---|
| MTN MoMo | Pre-approval | Yes, request-to-pay | 1.5 - 3.0% |
| M-Pesa | Standing order/M-Pesa Ratiba | STK push | 1.5 - 3.0% |
| Airtel Money | Limited by market | Yes, USSD/redirect | 1.5 - 2.5% |
| Flutterwave/Paystack | Yes, card + tokenised | Yes | 1.5 - 3.5% |
| Card (Visa/MC) | Yes, native | Rare | 2.5 - 3.5% |
Rule of thumb: offer the token first where the rail allows it, and keep push-payment dunning as a universal safety net.
Dunning cadence and churn impact
A failed payment is not a lost customer: it is a customer to recover intelligently. The retry window and grace period make all the difference.
| Day after failure | Action | Channel | Cumulative recovery rate |
|---|---|---|---|
| D+0 | Silent automatic retry | Wallet API | 8 - 12% |
| D+1 | Reminder + payment link | SMS + push | 15 - 20% |
| D+3 | Soft-urgency reminder | 25 - 30% | |
| D+5 | Retry + email | Email + push | 30 - 35% |
| D+7 | Final notice before suspension | SMS + call | 35 - 40% |
| D+10 | Suspension + read-only grace | In-app | Final retention |
With a 7 to 10 day grace period and clean proration on plan changes, you avoid suspending customers who were simply going to pay two days late.
Monthly cost for 100 subscribers
| Item (2026 estimate) | Assumption | Monthly cost |
|---|---|---|
| Transaction fees | 100 x 6,000 FCFA (~9 EUR) x 2% | 12,000 FCFA |
| Failures to recover | 12% x 100 = 12 customers | - |
| Loss without dunning | 12 x 6,000 FCFA | 72,000 FCFA |
| Loss with 35% dunning | 12 x 0.65 x 6,000 FCFA | 46,800 FCFA |
| Net gain from dunning | Difference | +25,200 FCFA/month |
Dunning is not a luxury: across 100 subscribers it returns over 25,000 FCFA per month, roughly 300,000 FCFA per year.
Mini case study
Need a professional website?
Kolonell builds websites that attract clients, optimized for the Sénégalese market. Free quote in 2 minutes.
Brian runs InvoiceLite, a billing SaaS at 6,000 FCFA/month (~9 EUR) with 420 subscribers in Nairobi. His involuntary churn ran at 13%, about 55 subscribers lost each month. By enabling D+0 auto-retry, a D+3 WhatsApp reminder and an 8-day grace period, he recovers 36% of failures: 20 subscribers saved monthly. At 6,000 FCFA that is 120,000 FCFA/month retained, at near-zero dunning cost (WhatsApp + SMS). Over a year: more than 1,400,000 FCFA in revenue that would otherwise have evaporated.
FAQ
Can you really automate a monthly charge on MoMo or M-Pesa?
Partially. Depending on the market and business API, mandates/tokens (pre-approval, M-Pesa Ratiba) exist but remain less universal than cards. The most reliable 2026 model combines tokens where possible with push-payment dunning, at 1.5 to 3% per charge.
What is involuntary churn and why is it so high?
It is subscriber loss not by choice but by failed payment (insufficient balance, unfunded wallet, expired session). It runs 8 to 15% per month across the region, often higher than voluntary churn.
How long should an account stay in grace before suspension?
A 7 to 10 day window is the optimal order of magnitude: long enough to let the customer pay, short enough to limit free usage. Then switch to read-only rather than a hard cut-off.
Is WhatsApp dunning really more effective than email?
Yes, by a wide margin. On our 2026 estimates a D+3 WhatsApp reminder recovers 25 to 30% of failures, whereas email alone caps around 10 to 15%. WhatsApp open rates exceed 90%.
How do you handle proration on a mid-month upgrade?
Prorate by remaining days: moving from 6,000 to 15,000 FCFA on the 15th charges roughly half the difference, about ~4,500 FCFA immediately, then the full rate on the next cycle.
Let's talk about your project. We build your MoMo/M-Pesa subscription engine with dunning and proration built in. 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.
