The client
A licensed payment service provider processing card, bank transfer and wallet payments for more than 4,000 merchants, with volumes growing at over 8% a month.
The challenge
Reconciliation was handled by an operations team of 14 using spreadsheets and database queries. Each morning, the team downloaded processor reports, bank statements and the internal transaction ledger, then matched them manually.
As volumes passed one million transactions a day, the process broke down:
- Late settlement. Merchant settlements were regularly delayed until the afternoon, generating complaints and support calls.
- Growing suspense balances. Unmatched items accumulated in suspense accounts, and older items became hard to investigate.
- Revenue leakage. Fees on some transaction types were not being billed correctly because pricing rules were applied manually.
- Audit concerns. The external auditors and the regulator both raised the size and age of unreconciled items.
What we deployed
We implemented our reconciliation and settlement platform, integrated with the client's processors, banks and ledger:
- Automated ingestion of processor reports, switch files, bank statements and wallet ledgers as soon as they arrive.
- Three-way matching between processor, bank and internal ledger, with rules for partial captures, reversals, refunds and chargebacks.
- Settlement engine that calculates each merchant's gross, fees, reserves and net settlement automatically.
- Pricing engine applying each merchant's contracted fees, including tiered and capped pricing.
- Exception workflow routing unmatched items to the right team by category, with ageing and service levels.
- Daily reconciliation pack emailed to the head of operations, finance director and compliance officer at 07:30.
How we delivered
The first payment rail (card transactions) went live after two weeks, running in parallel with the manual process for five days. Bank transfers and wallets followed over the next six weeks. Historical suspense items were loaded and worked through using the exception workflow.
Results after six months
| Measure | Before | After |
|---|---|---|
| Auto-match rate | Not measured, mostly manual | 99.6% |
| Merchant settlement completed | Often after 14:00 | Before 09:00 every working day |
| Unreconciled items older than 30 days | Large and growing | Reduced by 94% |
| Fee billing accuracy | Manual, with gaps | Automated, 3.2% increase in fee revenue |
| Operations staff on reconciliation | 14 | 4, with 10 redeployed to merchant support and risk |
Reporting that changed decisions
The daily reconciliation pack shows match rates by rail and processor, open exceptions by category and age, suspense balances and settlement status. Within weeks it revealed that one processor's reports contained recurring timing differences that inflated exceptions. The issue was resolved with the processor, removing thousands of exceptions a day.
The fee accuracy report identified merchants who had been undercharged since onboarding because their pricing was entered incorrectly. Correcting the pricing recovered 3.2% of fee revenue going forward.
The compliance officer now presents reconciliation metrics to the board risk committee monthly, and the regulator's earlier finding was closed.
What made the difference
- Automation at scale. Rule-based matching handles volumes that no manual team could keep up with.
- Exceptions, not everything. Staff focus only on the 0.4% of transactions that need attention.
- Pricing in the system. Fees are calculated from contracts, not from memory or manual updates.
- Evidence for regulators. Reconciliation status and history are always available for audit and inspection.
Client identity withheld under a confidentiality agreement. Figures are drawn from the client's before-and-after reporting for the engagement period; outcomes for your business will depend on your starting point and scope.