Reconciliation is one of the first areas where things start to feel off after a treasury system goes live.
During testing, matching rules appear to work as expected. Transactions reconcile, reports balance and outputs look consistent.
Once the system is used day to day, that behaviour often changes.
Reconciliations that previously matched begin to produce exceptions. Items that were expected to clear remain open. Small differences start to appear across cash, accounting or reporting outputs.
Why this happens
Testing environments are controlled. Transaction volumes are lower, data is cleaner and scenarios are predictable.
In live environments, that consistency disappears. Bank data varies, descriptions change and timing differences between systems become more visible.
Where the gaps usually sit
- Differences in bank transaction descriptions and references
- Variations in file formats across banks
- Timing differences between systems
- Static data inconsistencies
- Configuration decisions that do not hold in daily use
What this impacts
When reconciliation becomes less predictable, the impact is not just operational.
Confidence in the numbers starts to drop. Teams spend more time investigating differences. Manual adjustments increase. Reporting needs extra checks. Questions come up around whether issues are data-related, configuration-related or process-related.
This creates friction across treasury, finance and technology teams.
What to check
When reconciliation behaviour starts to drift, it helps to step back and review:
- How matching rules were defined and what scenarios they were tested against
- Whether bank data structures and descriptions vary more than expected
- How timing differences affect when transactions are available for matching
- Whether static data and reference fields are consistent across processes
- How exceptions are handled and whether patterns are emerging
This is less about fixing individual breaks, and more about understanding how the system behaves in live use.
Closing
Reconciliation working in testing does not guarantee it will behave the same way after go-live.
The difference is not always in the rules themselves, but in the environment those rules operate in.
Understanding that difference helps teams move from reacting to issues, to managing them more effectively.
Related insights:
- Why treasury system issues often appear after implementation
- When treasury system issues occur, the root cause is often data
- Why treasury system reconciliations don’t match after go-live
If you're working with treasury systems in live environments, you'll know that many issues only become visible after go-live.
I've covered these patterns in more detail, including reconciliation behaviour, system design gaps and control risks.