How to Audit Timesheet Rounding for Compliance

A step-by-step method to test whether a rounding policy is neutral over time, with a worked audit and a downloadable approach.

Actualizado 2026-10-20

Esta guía está disponible en inglés por ahora; la traducción al español llegará pronto.

Rounding is legal only when it does not systematically short employees. The only way to know is to measure it. This guide shows how.

What you need

Raw punch data: employee, date, in time, out time, ideally unrounded to the minute. Most time-clock systems export this as CSV. You also need the rounding rule in use (nearest 15, 6 or 5 minutes) and the direction policy.

Step 1: Compute exact worked time

For each shift, exact minutes = out − in (adding 24 hours across midnight) minus unpaid breaks. Sum per employee per pay period.

Step 2: Compute rounded worked time

Round each punch to the policy interval with the half point going up, then subtract. Sum the same way. Do not round the daily totals.

Step 3: Compare

Rounded minus exact, per employee per period. A neutral policy produces differences that scatter around zero and average close to zero across many periods. A policy that is losing employees time shows consistently negative differences.

Worked audit

A three-minute gain in the employee’s favour over the week is consistent with a neutral policy. Over a year, look for a cumulative drift larger than a few minutes per week.

Red flags

• Differences are negative in almost every period.
• Clock-ins round up and clock-outs round down.
• Grace periods: employees told to clock in early but paid from the scheduled start.
• Overtime thresholds crossed only in exact time, never in rounded time.

Fixing problems

Switch to exact time (most systems support it), or at minimum change to nearest-interval rounding on both punches. Back pay for the measured shortfall, plus liquidated damages in federal claims, is the usual remedy, so fixing it early is far cheaper.

Using the audit tool

Paste your export into the Timesheet Rounding Audit, pick the interval, and it produces the table above for every employee and period, with CSV export for your records.

Sample size and statistics

A single week can mislead; rounding differences are noisy. Audit at least a quarter, preferably a year, and look at both the average difference per shift and the total per employee. With nearest-interval rounding on both punches, the expected average is zero and the standard deviation per shift is roughly a third of the interval (about 4 to 5 minutes for quarter-hour rounding). An average that sits more than a minute or two below zero across hundreds of shifts is statistically unlikely to be chance and is the signature of a biased practice or of employees systematically arriving just after the round-down boundary.

Separating policy bias from behaviour bias

Rounding can short employees for two reasons: the rule itself is biased (round-down only), or the rule is neutral but behaviour is not, such as employees being told to clock in at 7:53 for an 8:00 start and clock out at 4:34. The second is harder to see in the numbers and easier to see in the distribution of punch minutes. Plot how many punches fall in each minute of the quarter hour; a neutral environment is flat, a managed one piles up just after the boundaries that round down.

Overtime effects

Rounding can also move hours across the 40-hour line. An employee at 40:05 exact and 40:00 rounded loses five minutes of overtime, worth more than five minutes of straight time. Flag every shift where the exact total crosses an overtime threshold and the rounded total does not.

Record keeping

Keep the raw punch export, the audit workbook and a note of the method for at least three years, the FLSA retention period for payroll records. If you change the policy as a result, date the change and keep the old audit.

Frequently asked

Our system rounds in but not out. Is that ever acceptable? No. Rounding only clock-ins up and clock-outs down is the textbook example of a non-neutral practice.

We round to 15 minutes but pay to the minute for overtime. Is that neutral? Mixing methods complicates the analysis; audit both the straight-time and overtime effect separately.

How much back pay does a shortfall create? The measured minutes at the regular rate, at 1.5x where they would have been overtime, for every affected employee over the look-back period, plus liquidated damages equal to that amount in federal claims unless the employer shows good faith.

Is exact-time payment safer? Yes, and most modern systems support it with no extra cost. The audit tool is just as useful afterwards to prove the change worked.