The tracking of employee working hours, breaks, and absences.
Time and attendance is the recording of when employees start work, stop work, take breaks and are absent, producing the hours-worked figure used to pay them and to prove working-time compliance. It is an operational tool and a pay record at the same time, so the numbers it produces have to survive scrutiny from a worker, an auditor or a tribunal.
At minimum the record shows, per worker per day: the start of work, the end of work, unpaid breaks taken, and a reason code for any hours not worked - sickness, annual leave, training, lateness. Mature records add three things that matter more than the raw punches: the shift the day was measured against, so a variance is visible; the exception and its approval, so unplanned overtime is authorised rather than accrued silently; and an edit trail showing who changed a punch, when and why.
That last item is what separates a record from a list of numbers. A messy record with provenance can be explained. A tidy record that somebody retyped, with no trace of the original reading, cannot - and the party who has to explain it is the employer.
Two duties turn attendance data into evidence. The first is the working-time limit: a worker "cannot work more than 48 hours a week on average", normally averaged over 17 weeks. Averaging is the operative word. You cannot demonstrate compliance from this week's timesheet alone; you need the preceding weeks intact, per worker, and you need to know whether that person has opted out, because a worker may opt out of the 48-hour week. Workers under 18 "cannot work more than 8 hours a day or 40 hours a week", so the record must also know who is a young worker.
The second duty is the payslip. Payslips must be provided on or before payday, and must show earnings before and after any deductions, the amount of any deductions that may change each time someone is paid - tax and National Insurance, for example - and "the number of hours you worked, if your pay varies depending on time worked". Where pay is hourly or variable, the attendance figure is not an internal metric. It is published to the worker every pay run. If the timesheet and the payslip disagree, the discrepancy has been handed to the worker in writing.
Most systems round clock times onto a fixed interval. Rounding is not inherently a problem; asymmetry is. If arrival is rounded forward to the next interval and departure is rounded back to the previous one, every shift loses a few minutes and the loss only ever runs one way. Across a workforce and a year that is a systematic reduction in recorded, and therefore paid, hours - and it is trivially provable from the unrounded punches the same system kept.
Two tests keep the practice defensible. Symmetry: rounding should be as capable of adding time as removing it, and over a pay period should net close to nothing for each worker. Separation: a grace period is a conduct decision about whether someone is treated as late, while rounding is a pay decision about how much time is recorded. Conflating the two produces the familiar dispute in which a worker is counted present from the scheduled start but paid from a rounded one, or is required to be on site for time the record never captured.
Attendance is not scheduling and it is not productivity. A rota says what was planned, attendance says what happened, and neither says what was achieved. Systems that blur them end up reporting planned hours as worked hours, which quietly makes the record untrue for everyone whose day departed from the plan.
The number that misleads most often is the team average. A department averaging comfortable weekly hours can still contain individuals over the limit, because the duty attaches to each worker's own averaged hours rather than to a headcount mean. Attendance rates behave the same way: a healthy-looking percentage can sit on top of a few people absent repeatedly, which is a different problem with a different remedy. Report per-worker rolling figures and distributions; the departmental average is what gets quoted upwards and what fails first when somebody checks it.
It must show "the number of hours you worked, if your pay varies depending on time worked" - so hourly and variable-hours staff have to see the hours their pay was calculated from. Payslips must be provided on or before payday, and must also show earnings before and after any deductions, plus the amount of any deduction that can change each time someone is paid, such as tax and National Insurance. Deductions that are fixed in amount must be explained either on the payslip or in a separate written statement. In practice this means the attendance figure and the payroll figure have to reconcile before payday, not after it.
On average rather than week by week: a worker "cannot work more than 48 hours a week on average", normally averaged over 17 weeks. One long week is therefore not automatically a breach, and a quiet week can offset a busy one - but you can only show that if the record retains the whole reference period for that individual instead of just the current week. A worker may opt out of the 48-hour week, so opt-out status is a per-person fact the record has to carry. Workers under 18 are treated differently: they "cannot work more than 8 hours a day or 40 hours a week".
Rounding to a fixed interval is common practice and is not the issue in itself; one-directional rounding is. If arrival is always rounded forward and departure always rounded back, recorded hours fall short of hours actually worked on every single shift, and the unrounded punches held in the same system make that pattern easy to demonstrate. Keep rounding symmetrical, retain the raw times, and be able to show that over a pay period the adjustment nets close to zero for each worker rather than running consistently in the employer's favour.
Explore our full platform with a 14-day free trial on Standard. Manage employees, post jobs, hire faster, and manage your tasks effortlessly with our all-in-one platform.