The Numbers Worth Watching

A site access system produces a great deal of countable material, most of which describes how big the project is.

For teams turning the numbers worth watching into a repeatable record, read more provides a useful operational reference for time, attendance and reporting workflows. It should be assessed alongside ISO standards catalogue so the software choice remains grounded in site practice and applicable guidance.

The seven

Unmatched entries, daily, as a proportion.

Time to produce the evacuation register, from the last drill.

Drill discrepancies, split into data errors and people.

Competence expiries already past, and those due in thirty days.

Temporary accesses granted, and how many are still open.

Access refusals, by reason.

And people on site whose engaging company cannot be identified.

All seven describe the arrangement rather than the workforce. That is deliberate.

What each says when it moves

Unmatched entries rising: the exit arrangement has degraded, or a new gate opened without a reader.

Drill time rising: the register is slower to produce, usually because it moved to a system that needs a login.

Expiries accumulating: nobody is running the forward list, which means the owner has stopped or never started.

Temporary accesses rising: enrolment has stalled, or programme pressure has arrived.

Refusals falling to zero: the enforcement has quietly been turned off, which is more likely than everybody suddenly complying.

That last one is the counter-intuitive read and it is worth checking rather than celebrating.

What not to measure

Individual attendance for people you do not employ, which is somebody else's management and will produce a complaint rather than information.

Hours per person as a productivity figure.

Headcount as a performance measure, since a subcontractor with more people on site is frequently one that is behind.

And anything that makes an individual's use of the gate a thing they are judged on, which produces workarounds and a register that describes less.

Who gets what

The register owner: all seven, weekly.

The site manager: drill result, temporary accesses, refusals, and the welfare comparison.

The project team: the pressure indicators — temporary accesses, hours exceedances, open permits — alongside progress.

The client: whatever the contract requires, produced on schedule rather than on request.

And the subcontractors: their own reconciliation and their own outstanding evidence, which converts chasing into a routine.

The comparison that matters most

Register against a physical head count, unannounced, at two in the afternoon.

One number, ten minutes, and it validates or destroys everything else.

Run it monthly. A site that has never run it does not know whether its register is a record or a list.

The annual or project-end view

Did the drill time improve?

Were the expiries prevented rather than caught?

Did the chain data stay complete as the project deepened?

Was the archive completed?

And one honest statement about what was not achieved, which is what makes the rest of the report credible.

The measure behind the measures

Whether anybody acts on them.

A set of figures produced weekly and read by nobody is an administrative artefact, and the test is simple: when a number last moved, did anything change?

If the answer is no for six months, the problem is the ownership rather than the measurement.

Producing them without a system

All seven come from records a site already holds.

Unmatched entries from the access export.

Drill time from a stopwatch.

Expiries from a spreadsheet of dates.

Temporary accesses from the authorisation log.

None of this needs a dashboard, and a site waiting for one measures nothing for a year.

Reporting to a client

One page: the drill result, the register accuracy check, the expiry position, and what changed as a result.

Plus one honest statement about what has not been achieved, which is what makes the rest credible and is the section most often omitted.

The one to publish to the workforce

The drill result.

Because it is the only number that tells people the register exists for them rather than about them, and publishing it commits the site to improving it.

Most sites publish nothing, which leaves the workforce to infer the purpose from the hardware — and the inference is always the least favourable one.