Research note
How often do timekeeping imports disagree with the approved schedule?
A quarterly method for comparing a frozen timekeeping import with the approved schedule and classifying each disagreement by cause.
Published

Table of contents
Question and scope
The question covers employees whose hours are imported from a timekeeping system into payroll. It includes regular hours, overtime, premium time, and approved unpaid breaks. It excludes manual payments and one-time adjustments that never pass through the import. The scope should be written before the sample is drawn.
An employee-period pair is eligible when the employee appears in both the import and the approved schedule. Pairs that appear in only one source are recorded separately as coverage gaps. Record the period, the schedule version, and the import file. This keeps a missing employee from being counted as a zero-hour match.
The review also records whether an approved correction was applied after the import. A correction can make a correct import look wrong if the reviewer compares it with the final pay instead of the import. Use the frozen import as the comparison point. Report corrections separately.
Definitions and evidence standard
Define a disagreement as a difference in hours, rate, or rule application that the approved schedule or policy does not explain. A rounding difference within the stated rule is not a disagreement. A missing approved exception is a disagreement even when the total happens to match. The definition should be fixed before the review.
Use the frozen import, the approved schedule, and the policy version as evidence. The reviewer should be able to reproduce each value from those sources. Record the person who checked the case and the date. A difference without a source reference is only an impression.
Separate a data disagreement from a rule disagreement. A wrong employee identifier is a data problem, while a correctly captured hour with the wrong overtime rule is a rule problem. The repairs are different. Label the case before assigning an owner.
Sampling and measurement
Freeze the import and the schedule at the same time and record both export times. Include more than one location, shift pattern, and employee type. Select the sample by employee identifier rather than by name. A sample drawn from the largest department alone will miss boundary cases.
Measure the number of pairs reviewed, the number with a disagreement, and the number with an unexplained difference. Add the number of coverage gaps and the number of corrections made after the import. Report the counts by disagreement type. A single error rate hides whether the problem is data or rules.
Check the first and last day of the period in every sampled pair. Boundary errors are common and easy to miss in a total. Record the scheduled hours and the imported hours for those days separately. The extra detail shows whether the disagreement is a boundary issue.
Findings and interpretation
Present the disagreement rate by type before suggesting a cause. A cluster of identifier mismatches may point to a mapping table that was not updated. A cluster of rule differences may point to a policy change that the import did not receive. Read the cases behind the largest cluster before naming a reason.
Compare pairs with a schedule change during the period against pairs with a stable schedule. A change can create a valid split that looks like an error in a simple total. Report the change date and the split. The comparison keeps a correct split from being counted as a defect.
Check whether the disagreements were corrected before pay. A pre-pay correction may protect the employee even when the import failed. Report the correction date and the reviewer. The timing matters for the control, not only for the final amount.
Limitations
The review measures the import against the approved schedule, not against the hours actually worked. A schedule can be wrong while the import matches it. It also cannot show whether the employee was paid correctly after every adjustment. The result describes one control point.
The schedule may change during the period, and the version used for the review may not match the version in force at the time. Record the effective date rather than applying the final version retroactively. A difference under an old schedule may not be a current defect. The reviewer should state which version applied.
Deciding whether a difference is material requires judgment. A small rounding difference and a missing overtime premium are not the same. Record the threshold and the reason. A calibration set helps reviewers apply the same standard.
Practical next step
Add a pre-import check that compares the employee identifiers in both sources. The check should stop the import when an identifier is missing or duplicated. Record the result with the import file. A short check before the run is easier than a repair after pay.
Assign one owner to the mapping table and one reviewer to the rule application. The owner can update the mapping when a system changes, while the reviewer can confirm the policy. A shared ticket is not a substitute for a named owner. Review both after any system change.
Run the same spot check next quarter and compare the disagreement rate with the baseline. Keep the pre-import check if it catches a gap before pay. If the rate does not move, inspect the source schedule before adding more review. The goal is a clean import, not a longer exception list.
Sources
References supplied with this research note.
| Source | Publisher | Context |
|---|---|---|
| U.S. Department of Labor, Recordkeeping | dol.gov | Primary reference supplied with this article. |
| Internal Revenue Service, Publication 15 | irs.gov | Primary reference supplied with this article. |
| National Institute of Standards and Technology, Cybersecurity Framework 2.0 | nist.gov | Primary reference supplied with this article. |
| U.S. Department of Labor, Fair Labor Standards Act | dol.gov | Primary reference supplied with this article. |