What Not to Do

Each of these is common, understandable, and leaves the site worse than the arrangement it replaced.

For teams turning what not to do into a repeatable record, site-team productivity tracking provides a useful operational reference for time, attendance and reporting workflows. It should be assessed alongside Chartered Institute of Building so the software choice remains grounded in site practice and applicable guidance.

Automatic sign-out at a fixed time

Why not: it records an assumption as a fact. The register then looks complete and is unusable for the one purpose that cannot tolerate error.

Instead: make leaving easy, and report unmatched entries as a known gap rather than closing them artificially.

Treating the gate as a timesheet for other people's staff

Why not: it is somebody else's management, it will produce a complaint from the subcontractor rather than better attendance, and it teaches the workforce that the system is about them rather than about safety.

Instead: presence for safety, and a monthly commercial reconciliation shared with the subcontractor.

Biometrics because card sharing might happen

Why not: a substantially higher legal bar for a problem that has not been measured, and consent that cannot be freely refused.

Instead: measure the sharing first. Turnstiles solve most of it.

Verifying competence only at induction

Why not: cards expire, and a person verified in March is unverified by September with nothing to show it.

Instead: expiry held and enforced at access, with a forward list.

Admitting an unverified person because the job is urgent

Why not: the urgency is permanent and the exception becomes the process within a month.

Instead: a written temporary access rule with an authoriser, an expiry, and a weekly report.

One register for access and plant

Why not: the gate either blocks everybody or verifies nothing.

Instead: two registers, separate owners, consolidated weekly.

Collecting documents because they might be useful

Why not: the site ends up holding identity documents, qualifications and medical information it cannot justify, about people it does not employ.

Instead: status and expiry, an emergency contact, and the right-to-work check recorded rather than copied where the rule allows.

Leaving the archive to project completion

Why not: at handover everybody is leaving and nobody owns it, and the questions arrive years later.

Instead: decide the archive at project start with a named owner.

Buying a system before deciding who owns the register

Why not: every failure in this collection follows from an unowned register, and the system will not create the owner.

Instead: name the person first. They will specify the system better than a procurement exercise will.

The common thread

Most of these substitute a mechanism for a decision — about staffing the gate, about what the register is for, about who maintains it.

The mechanism is purchasable and visible. The decision is neither, which is why the substitution keeps happening and why every correction above costs nothing.

Three more

Announcing a deadline for enrolment and moving it. The last wave then never completes, and the exceptions become the arrangement.

Running the system without an evacuation drill. Months of confident data with the one capability untested.

Letting the subcontractor hold the only record of who they brought. It is accurate and it is not available at seven in the morning when somebody is missing.

How to tell which you have

Ask five questions and watch one thing.

When was the last drill, and what was the time to produce the register?

How many unmatched entries yesterday?

How many temporary accesses this month?

Who owns the register by name?

And stand at the gate at finishing time.

The recovery

A site that already has all of these is not in a bad position.

Name an owner, run a drill, fix the exit, and publish the unmatched-entry rate.

One month, no purchase, and the register becomes something people trust rather than something they route around.

Why the visible thing wins

A turnstile can be photographed for a board report. A named owner cannot.

A system has a price and a delivery date. A decision about who maintains the register has neither.

Which is the mechanism behind every item above, and naming it is the only defence: when a proposal in this area arrives, ask which decision it is standing in for.