The Gate

Everything in this subject depends on knowing who came through the gate and who went back out. The methods differ mainly in how they fail.

For teams turning the gate into a repeatable record, Monitask 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.

The sign-in book

A paper register at a gate or a cabin.

Cheap, works in a power cut, and requires nobody to be trained.

Fails at the exit, which is the whole problem: people sign in because somebody is watching and leave without signing out because nobody is.

And it cannot be searched. A roll call from a paper book on a wet day with three hundred names is not a roll call.

Adequate for a small site with one entrance and a gateman who knows everybody. Nothing larger.

Card and fob readers

A card presented at a reader, in and out.

Searchable, timestamped, and the register is live.

Fails through card sharing, which happens when somebody forgot theirs and the alternative is losing a day's pay.

And through tailgating: one card, two people through a door.

Which makes the reader adequate where the incentive to share is low and useless where it is high — a distinction worth making honestly at the design stage rather than discovering later.

Turnstiles

A physical barrier that admits one person per authorisation.

Which is the only arrangement that actually stops tailgating, and that is what you are buying.

Costs installation, a power supply and a layout that can accommodate them.

And produces a queue at shift start, which has to be designed for: too few lanes and the site's first act each morning is a twenty-minute wait in the rain.

Biometric readers

Fingerprint or facial, tying the record to the person rather than the card.

Which resolves sharing entirely.

And introduces biometric data with a substantially higher legal bar, consent that cannot be freely given in an employment context, and a workforce reaction that is frequently out of proportion to the problem being solved.

Reserve it for the case where sharing is a demonstrated problem and a lesser method has been tried, which is the proportionality argument you will need anyway.

The failure that defeats all of them

Nobody records leaving.

Every method above is accurate on entry and depends on cooperation at exit, because there is no benefit to the individual in signing out and frequently a queue in the way.

Which means the register drifts upward through the day, and by mid-afternoon it describes who arrived rather than who is present.

The roll call then produces a list of people to search for who went home at lunchtime, which is worse than useless in an emergency: it consumes the marshals' attention at the moment it is scarcest.

Designing the exit

Make leaving as easy as arriving, which usually means the same number of lanes in both directions.

Automatic sign-out at a stated time is a common compromise and should be understood for what it is: an assumption recorded as a fact, and unusable as evacuation evidence.

Better: a sign-out prompt tied to something people want, which on many sites is the vehicle barrier or the welfare facility.

And measure the gap: entries without a matching exit, per day, as a proportion. It is the single most useful number about the gate and almost nobody produces it.

Multiple gates and vehicle access

Every additional entrance multiplies the problem.

Vehicle gates in particular: people arrive in a van, and only the driver presents a card.

Either the vehicle gate has a reader for passengers or the site accepts that the register is wrong by however many people came in the back of vans.

Say which, because an unstated compromise becomes an assumed capability.

Visitors and deliveries

A different flow with the same requirement: a driver delivering blocks at ten and leaving at ten past is on site, and would need finding in an evacuation.

Short-duration visitors are where registers are weakest and where the effort to fix them is least justified operationally and most justified in an emergency.

A simple, fast visitor process that produces a timestamped record is enough. The failure is having no record at all, not having an imperfect one.