What the Records Cannot Tell You
Questions routinely asked of attendance data that it cannot answer, and where the real answer is.
Reference · Analysis
Timestamp data is precise and narrow. The gap between those two properties is where most of the trouble in this subject sits.
Whether someone was working
It records presence, or the presentation of a credential.
Not effort, output or attention.
Where to look instead: the work.
Whether an anomaly was deliberate
Six of the common causes are system failures or structural arrangements.
A timestamp cannot distinguish them from a choice.
Where to look instead: the site's anomaly rate, the system status, and the person's own account.
Who presented the credential
A card, a code and a phone are all transferable.
Only identity verification addresses this, and the record alone never does.
How much time theft you have
The category is six things, and you can measure two of them.
Any total is therefore partly invented, and a total presented as a fact is the thing that gets challenged first.
Where to look instead: measure the two that are measurable and say the others are estimates.
Whether the workforce is honest
It is not a question this data answers, and framing it that way changes what people do.
Where to look instead: the pattern. If it follows a site, a shift or a rota change, it is not about honesty.
Whether the policy is working
A fall in recorded anomalies can mean the causes were fixed or that people stopped reporting failures and started absorbing the minutes.
Where to look instead: corrections submitted, and asking people.
What it does support
That a credential was presented at a time and place.
That a shift was recorded as lasting a certain duration.
That corrections were or were not made, by whom.
That rounding moved minutes in a measurable direction.
Four things, recorded consistently, and the fourth is the one most organisations have never looked at.
Say so when asked
The discipline that keeps the rest credible.
When asked for a total the data cannot support, saying so is the answer.
A number produced anyway circulates without its caveat and returns as an expectation.
In this subject it also ends up in a disciplinary meeting, where the person is entitled to ask how it was calculated.
Name what the records do support and stop there.
Reproduce the workflow
For another way to make this requirement testable, consult this sector page. Reproduce the case with real roles, codes, failures and recovery steps.