Lone working in security: what good practice actually looks like

A large share of contracted guarding is, by its nature, one person on a site with nobody else around — overnight static posts especially. Lone working is not something to eliminate from the model; it is central to it. What varies enormously between contractors is whether there is an actual plan for what happens if that officer needs help.

The general duty, and the sector reality

Employers have a general duty under health and safety law to assess and manage the risks of lone working, and the current HSE guidance is the place to check what that duty involves for your operation specifically. It is a generic duty applied to a sector where lone working is close to the default, which is exactly why it needs deliberate attention rather than being assumed to be covered by general policy.

What a lone working plan actually needs to cover

• –A defined check-in schedule, so an officer who misses one is noticed within a known window, not whenever someone happens to call
• –A clear escalation path if a check-in is missed — who is contacted, and in what order
• –A means of raising an alarm that does not depend on the officer being able to use a phone normally
• –Site-specific risk factors identified in advance — isolation, access, known local issues — rather than assessed only after something happens
• –A record of who is on a lone-working post on any given night, so an escalation actually knows where to start

The last point sounds basic, but it depends on the same information a rostering decision needs — see cover at 5am for how that information tends to live in the wrong place under time pressure. A lone-working plan is only as good as knowing, in real time, who is actually lone working tonight.

Where plans usually fail in practice

Not from the absence of a policy document — most contractors have one. They fail because the check-in schedule is not actually followed under pressure, because a device relied on for alarm-raising was not checked before the shift, or because the escalation contact list is out of date and the first call goes nowhere.

A policy that exists on paper and a plan that would actually function at 3am on an isolated site are not automatically the same thing, and it is worth testing which one you actually have.