A control total is a reference figure you calculate before transforming data, then compare with the result afterward. For a timesheet import, a matching grand total is useful but insufficient: two employees' hours could be swapped while the overall total stays unchanged. Check counts, people, dates and hours together.
Establish the source totals
Start with an approved source export and keep an unchanged copy. Record the number of rows, unique employee identifiers, earliest date, latest date and total duration. Convert hours and minutes into a single exact unit, such as whole minutes, when the source precision permits it.
For example, three records of 7:30, 8:15 and 6:45 contain 450, 495 and 405 minutes. The reference total is 1,350 minutes, or 22.5 hours. If a conversion produces 22.30, someone probably combined the display digits rather than converting the minutes.
| Check | Before conversion | After conversion |
|---|---|---|
| Rows | 3 | 3 |
| Total minutes | 1,350 | 1,350 |
| Total decimal hours | 22.5 | 22.5 |
| Rejected rows | Not applicable | 0 |
Compare at more than one level
Create totals by employee and by date as well as for the full file. An eight-hour row moved from employee A to employee B leaves the grand total unchanged but breaks both employee totals. A row assigned to the wrong week may leave an employee's monthly total correct while making the weekly reports wrong.
Where separate codes represent work, leave or adjustments, reconcile those categories too. Equal combined hours do not prove equal classification. Keep identifiers as text when leading zeros matter, and compare the original identifier to the destination value rather than relying only on display names.
Investigate differences rather than forcing a match
A mismatch can come from dropped blank rows, duplicate records, filtered data, text values ignored by a sum, date interpretation or rounding. Check the import's rejection report before changing the source. A rejected row may explain the gap without any arithmetic error.
Do not add a balancing row simply to make the total agree. Find the underlying record, correct the mapping or source with an audit note, and rerun the comparison. If the target intentionally rounds at a different stage, calculate the expected effect separately and record the rule used.
Save a compact sign-off record
Keep the source filename, destination batch identifier, conversion rule, counts, totals, exception list and review date. State whether the figures describe hours worked, paid hours or another measure. Someone reviewing the import should be able to reproduce the comparison without reconstructing assumptions from memory.
A control total validates transfer consistency. It does not prove that the original time entries were accurate or approved. Those checks belong upstream. The final record should distinguish “matches the approved source” from “independently verified against clock events.”
Frequently asked questions
Can a matching grand total hide missing rows?
Yes. A missing positive entry and an extra entry of the same size can cancel out. That is why row counts and employee-level checks matter.
What tolerance should I use?
Use the documented precision of the workflow. For exact whole-minute data with no intended rounding, require an exact minute match. Explain any permitted tolerance instead of treating unexplained differences as harmless.
Related calculators and guides
- Decimal Hours Calculator & Bulk Timesheet Converter
- Check a Decimal-Time CSV Before a Payroll Import
- Seven Common Timesheet Errors and How to Catch Them
Explore more in Hours & Pay Calculators and Guides.