The verdict in three sentences
Manual reconciliation between the M-Pesa statement and your orders is slow, painful and riddled with errors: 3 to 6% of lines do not match on the first pass. By automating matching on the merchant reference and gap detection, you cut the time fivefold and make every unit traceable. The result: reliable accounting, a controlled T+1 settlement and zero silent gaps.
What reconciliation actually catches
Reconciliation means confronting three sources: your order database, the operator statement from the Daraja API (CSV export) and your accounting. Gaps almost always come from the same causes.
| Gap type | Frequency (2026 estimate) | Typical cause |
|---|---|---|
| Unrecorded operator fees | 1.8% of amount | Commission taken at source |
| Duplicate transactions | 0.5 to 1% | Replayed webhook, customer double-click |
| Unmatched refunds | 0.8% | Product return not linked to order |
| Payment received with no order | 0.3% | Wrong merchant reference |
| Paid order not credited | 0.4% | Late settlement, dispute |
Combined, these anomalies represent 3 to 6% of lines to handle by hand every month. At volume, that is a full working day lost.
The four-step automation workflow
A good reconciliation pipeline always follows the same logic:
| Step | Action | Rule applied |
|---|---|---|
| 1. Import | Pull the M-Pesa CSV export | Daily, T+1 |
| 2. Matching | Cross by merchant reference + amount | 0 tolerance on the net |
| 3. Detection | Isolate unmatched gaps | Alert threshold > 500 FCFA |
| 4. Processing | Classify: fees, duplicate, refund | Automatic ledger entry |
Matching on a unique merchant reference is the key: each order generates an identifier that you inject into the payment reference field. Without it you are reduced to matching by amount and timestamp, which breaks the moment two customers pay the same sum in the same minute.
Handling fees and T+1 settlement
M-Pesa takes its commission at source: you sell at 10,000 FCFA but the net credited is lower. If your books record the gross, you will get a systematic gap of around 1.8%. The fix: record gross revenue and operator fees separately, using the fee line on the statement. Since settlement arrives at T+1, you always reconcile the previous day, never the current one.
Mini case study
Ibrahim runs an online grocery in Nairobi that collects 1,200 M-Pesa transactions per month for a total of 9,600,000 FCFA (about 5.9M KES equivalent). Manually he spends about 10 hours a month ticking off his statement, and lets 4% of lines slip through, meaning 48 transactions with small gaps.
Need a professional website?
Kolonell builds websites that attract clients, optimized for the Sénégalese market. Free quote in 2 minutes.
By applying automatic matching on the merchant reference and a 500 FCFA alert threshold, he only handles the real anomalies by hand, meaning fewer than 10 lines per month. Reconciliation time: 2 hours instead of 10, a 8-hour monthly saving. On the money side, he recovers an average fee gap of 1.8%, roughly 172,000 FCFA of correctly booked commissions and a finally accurate bottom line.
FAQ
How many lines fail to reconcile automatically?
Between 3% and 6% in manual reconciliation per our 2026 observations. With merchant-reference matching, you usually drop below 1% of real anomalies to handle.
Why does my revenue never match the statement exactly?
Because of operator fees taken at source, on the order of 1.8% of the amount. You must book gross and fees separately, otherwise the gap piles up every month.
What alert threshold should I set on gaps?
A threshold around 500 FCFA is a good compromise: below it, these are often fee rounding; above it, it warrants a human check.
How much time does automation really save?
About 8 hours per month on a volume of 1,000 to 1,500 transactions. That time is reinvested in customer service or sales.
Does T+1 settlement complicate reconciliation?
On the contrary, it structures it: you always reconcile the previous day, which gives a clear daily rhythm and avoids chasing payments still in transit.
Let's talk about your project. We automate your M-Pesa reconciliation with reference matching and a clean ledger. 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.
