The Checks Worth Running

None of these requires a purchase, and together they produce a picture of your own operation that no survey can.

Anomaly rate by site

Refused punches, missing clock-outs and corrections, as a proportion of shifts, per site, per week.

Segment by cause where the system records it.

A site far above the others has a fault, a queue or a changed entrance.

This is the headline number and most organisations do not have it.

Punch-to-shift distribution

How long before the shift people clock in, across everyone.

A tight cluster a few minutes early is the terminal location and the transport timetable.

A long tail is worth understanding before it is worth acting on.

One person far from everyone else is a different question.

Break distribution

Median and tail, not average.

A median of two minutes over with a long tail is a structural problem.

Everyone taking ten is a norm question.

Rounding neutrality

Aggregate raw against aggregate paid, over a quarter.

Near zero is neutral.

A consistent gap is the finding, and the direction tells you whose minutes.

Ten minutes, and it is the one most likely to produce a surprise.

Pre-shift time

Stand at the entrance at shift change and time ten people from arriving to being ready to work.

Multiply by headcount and working days.

Compare against whatever prompted the project.

Handover overlap

Rostered shift boundaries against actual punch times, by shift.

A consistent gap at every boundary is the structural overlap, and it has a value you can state.

Reporting them

All six, together, once a quarter.

Both directions on the same page.

By site rather than by person.

With the recording method stated, because a system that does not log failures produces an anomaly rate that understates itself.

What they replace

Industry percentages.

Vendor calculators.

And the argument about whether there is a problem, which becomes an argument about a number you produced yourself.

Put them in the calendar

The only reason any of them get done.

Six measurements, quarterly, attached to something that already happens.

Rounding neutrality after every system update as well.

An owner, or they belong to nobody.

And the results kept as a series, because a single figure says little and four quarters say a great deal.

Connect policy and data

The choices in this note can be compared with review this team workflow. Keep the written purpose in control and enable only the information needed at this boundary.