Missed punches and regularisation: a simple rule set for shift-based teams

A supervisor tells payroll that four people worked the Sunday shift and the machine did not record them. Payroll pays them. Two months later, during a review of overtime cost, somebody asks who authorised it. The supervisor remembers the conversation. Nobody else does, and there is no record of either the request or the approval.

Nothing improper happened. The four people almost certainly worked. But the business cannot demonstrate that, which means it also cannot demonstrate the times it did not happen — and that is the real cost of an undefined regularisation process.

Why punches get missed

It helps to separate the causes, because they need different responses and treating all of them as carelessness guarantees the rules will be resented and worked around.

• The device failed — power cut, network down, reader not recognising a worn finger on a cold morning. Common, and entirely the company's problem.
• The person was elsewhere — sent to another site, collecting material, at a customer. They worked; they were not at the reader.
• Enrolment was incomplete — a new joiner whose biometric was never registered, so a week of attendance simply does not exist.
• Shift boundaries — a night shift that crosses midnight and lands in the wrong day, or an early handover before the roster's official start.
• Genuine absence — the smallest category, and the one most rule sets are written as though they were about.

A rule set that acknowledges the first four is one supervisors will use. One that treats every missed punch as suspected absence pushes corrections back into conversations, which is exactly where you do not want them.

The four rules

1. 1

   A window

   Regularisation must be raised within a fixed number of working days of the date concerned, and in any case before the attendance cut-off for the month. Corrections after the month has closed are arrears, handled in the next cycle, not reopenings of a closed month.

2. 2

   An approver, by role

   The employee's own reporting manager, with a named deputy for when that person is unavailable. Not the supervisor who happened to be on the floor, and not HR — HR records the correction, it does not authorise the facts.

3. 3

   A required reason, from a short list

   Device failure, off-site work, enrolment gap, shift boundary, other. "Other" must carry a sentence. A free-text field with no list produces entries reading 'as discussed', which document nothing.

4. 4

   A cap that triggers review

   Above a set number of regularisations per person per month, or per supervisor per month, the case is reviewed rather than approved routinely. The cap is not a punishment; it is how you find the broken reader or the site that is never enrolled.

Those four turn regularisation from a favour into a transaction, which is what makes it defensible.

The record is the point

Whatever else you do, every regularisation should leave behind four facts: the date corrected, who requested it, who approved it, and the reason. Two months later that is the difference between a month you can explain and a month you cannot.

Employee / code            ______________________
Date being corrected       ____/____/________
Shift                      ______________________
Actual in / out claimed    ______  ______
Reason (from list)         device / off-site / enrolment /
                           shift boundary / other: __________
Raised by                  ______________  on ____/____
Approved by (role)         ______________  on ____/____

Cut-off for the month: ____/____   Raised after cut-off? Y / N
If Y: carried to next cycle as arrears. Month stays closed.

A paper form works. A form in the same system that holds attendance works better, because the correction sits against the day it corrects and cannot be separated from it later. That is what attendance, shift and leave work is ultimately for: every day of working time traceable to an approval with a name on it.

Use the pattern, do not just approve it

Regularisations are diagnostic data and are almost always discarded. Count them monthly by site, by supervisor and by reason, and they tell you things no other report will.

• One site producing most of the device-failure corrections has a reader that needs replacing, and the cost of replacing it is smaller than the cost of the corrections.
• One supervisor raising far more than the others is either running the only honest process or running a loose one, and the difference is worth a conversation.
• Enrolment-gap corrections concentrated in the first week of employment mean the joining process is not registering people — which belongs upstream in employee onboarding rather than here.
• Shift-boundary corrections clustered on one roster mean the roster is defined wrongly, not that people are arriving at odd times.

What to do about the month you cannot explain

Most businesses introducing these rules have a backlog: months already closed where corrections were made on somebody's word. Do not try to reconstruct them. Draw a line, start the rules from a stated date, and say so plainly to supervisors and employees.

Reconstruction rarely produces truth and always produces suspicion. What matters is that from the stated date forward, every corrected day has a name against it — and that the people raising corrections understand the rules exist so their requests are honoured routinely, not so they are refused.

The connection to payroll

One detail decides whether any of this holds: the regularisation window must close before the attendance cut-off, and the attendance cut-off must close before payroll opens. Where those dates are not sequenced, corrections keep arriving during the payroll run and the month never actually closes.

That sequencing is the same discipline that makes payroll process management work, and it is why the two are best fixed together. Attendance rules without a cut-off produce a payroll that is always being recalculated; a payroll calendar without regularisation rules produces a cut-off that everyone politely ignores.

What this looks like when it is working

Three signs, none of them dramatic. Supervisors raise regularisations routinely rather than mentioning them in passing, because the route is quick and the answer is usually yes. The monthly count is stable and explainable. And when somebody asks in March who authorised a Sunday in January, the answer takes a minute rather than a conversation.

There is a fourth sign that takes longer to appear. Arguments about attendance stop being about whether someone worked and start being about whether the rule is right — which is a far better argument to be having, and one that can actually be settled.

Who should not be approving regularisations

A detail that decides whether the whole rule set means anything: the person who approves a regularisation should not be the person who benefits from the shift being covered.

In practice, supervisors under pressure to show full coverage have an interest in attendance looking clean. A supervisor approving their own team's missed punches, while being measured on whether the shift ran, is being asked to hold two positions at once. Most will be scrupulous; the arrangement should not depend on it.

The fix is not to move approvals away from supervisors, who are the only people who actually know what happened. It is to make the record visible one level up, as a monthly count per team rather than as individual approvals to review. Nobody reads individual regularisations. A number that says this team regularised forty-one punches last month and the next team regularised six is read immediately.

That comparison is also the most useful diagnostic in the whole system. Wide variation between similar teams is rarely about employee behaviour. It is usually a device in a bad location, a shift pattern that does not match the punch windows, or a supervisor applying a different standard, and all three are fixable once the number makes them visible.

Businesses running multiple sites or shifts feel this most, because that is where the informal version breaks down fastest: the supervisor who knows everybody is not present at the second site, and the rule that lived in his judgement does not travel. Writing the rules down is what allows a second location to run the same way as the first, which is a large part of why attendance, shift and leave work tends to arrive on the agenda at the same time as expansion does.