The analytical claim

When two systems disagree, the age of the unresolved difference can reveal where operational records are losing reliability. It cannot, by itself, identify which system is correct. Treating every mismatch as an error in the downstream system is a common but unsupported shortcut.

NIST describes data integrity as protection against unauthorized or accidental alteration, corruption, and destruction, supported by asset awareness, integrity checks, logs, and recovery practices. The NIST Research Data Framework describes fitness for purpose through qualities including accuracy, completeness, consistency, timeliness, and provenance. Together, they support an analysis that preserves both versions and the history of the reconciliation decision.

Define a break before counting it

The unit is one expected correspondence between a source record and a destination record at a defined comparison time. Store the two stable identifiers, fields compared, extraction timestamps, normalization rule, difference class, opened time, investigation owner, disposition, and closure evidence. Never use a mutable customer name as the sole join key when a durable identifier exists.

Classify breaks before measuring them: missing source record, missing destination record, value difference, timing difference, duplicate candidate, mapping failure, or authorized exception. A timing difference may disappear after a normal synchronization window. An authorized exception may remain different by design. Keeping these classes separate prevents the backlog from becoming a misleading total.

Age begins when a break first exceeds the agreed synchronization or review window, not every time the report is rerun. Preserve the original opened time through reassignment. Resetting the clock makes old uncertainty look new and removes the very signal being studied.

Study design and measures

Run the same comparison at a stable cadence for a bounded period. Keep the matching and normalization rules versioned. Report the eligible population, successfully compared records, excluded records, new breaks, resolved breaks, reopened breaks, and open breaks by age band. The denominator matters: a low mismatch percentage is not reassuring if a large portion of records could not be compared.

Pair age with materiality and operational use. A two-day difference in a dormant reference field may be less consequential than a two-hour difference in a routing status used for same-day work. Record the decision or process that consumes each field, but leave business impact judgments to the accountable data and process owners.

Audit a sample of both matches and breaks against source evidence. Sampling only mismatches can reveal break causes but cannot detect false matches created by weak keys or normalization. Retain the original extracted values and reviewed conclusion. Corrections should append a reason and timestamp rather than erase the prior observation.

What the findings can support

A rising old-break share can indicate insufficient investigation capacity, a recurring mapping defect, unclear ownership, or unreliable source capture. A falling count may reflect genuine correction, a rule change, excluded records, or deletion. Show inflow, resolution, and rule changes together before claiming improvement.

An operations coordinator may schedule comparisons, maintain the break register, assemble supporting records, and route exceptions. The coordinator should not choose the system of record, merge possible duplicates, alter regulated data, or close a break without the named owner’s evidence.

Privacy controls are part of the method. Compare only the fields needed for the decision, restrict row-level exports, and avoid placing personal or confidential values in tickets. A stable record ID and controlled evidence link usually provide stronger traceability than copied sensitive data.

Conclusion

Reconciliation-break aging is a useful operational measure only when the comparison population, rule version, break class, and original open time remain visible. The strongest conclusion is not “system A is wrong.” It is that a defined difference has remained unresolved for a measured period and now requires a named decision.

OperationsCoordinator can support the recurring comparison, evidence collection, and owner follow-up through our operations support services. System-of-record and correction authority remain with the client’s accountable owners.

Sources and limits

These sources do not establish a universal break-aging limit or designate a source of truth for a business. This article is operational research guidance, not audit, accounting, privacy, legal, or cybersecurity advice.