How to Measure Payment Reconciliation: KPIs and Baselines
Build a retail reconciliation baseline with match rate, unresolved variance, exception age, and close effort before evaluating an integration repair.
Measure payment reconciliation with four indicators: match rate, unresolved variance, exception age, and close effort. Baseline them for a closed period before changing the workflow. Compare equivalent reporting periods after the change so shifts in channel or payment mix do not obscure the result.
Define the baseline
Choose one channel, payment provider, currency, and time window. Record the settlement cutoff, any periods still awaiting deposits, and the set of transaction types included. Keep exports or reports for the processor and ledger so another reviewer can reproduce the comparison.
| Indicator | How to calculate it | What to verify |
|---|---|---|
| Match rate | Matched eligible events ÷ total eligible events | Keep timing exclusions explicit and use the same denominator in both periods |
| Unresolved variance | Sum absolute unmatched differences by currency | Track net unexplained balance separately so opposite variances do not cancel |
| Exception age | Time from first detection to resolution | Report open and closed queues separately |
| Close effort | Staff time spent exporting, matching, reviewing exceptions, and correcting entries | Keep the same scope and roles in both periods |
Use the result to choose the first repair
If match rate is low because references are missing, fix the identifier mapping. If most remaining items are old, confirm who owns the exception queue and whether there is enough information to resolve items. If staff time remains high despite a good match rate, watch the actual close workflow and identify the manual checks that still repeat.
Validate before reporting a gain
Sample transactions labeled as correctly matched. Include refunds, split captures, and adjustments. Record changes to traffic mix and accounting rules alongside the numbers, and report any residual variance or manual work. A small before-and-after comparison is useful only when the same events and rules are measured on both sides.
See the implementation checklist for payment reconciliation and the scope of a first retail repair.
Next step
Need help applying this pattern?
Bring the bottleneck and the stack. We will map the first implementation path.