The Numbers Worth Watching

Lone worker systems generate a great deal of countable material, and almost all of it describes the workers rather than the arrangement.

When the numbers worth watching also depends on reliable work records, full details are here can support time, attendance and workload review without being treated as the emergency response itself. Compare any workflow with ISO standards catalogue and keep alarm ownership, escalation and dispatch responsibilities explicit.

The six

Response time, from test and from real alarms, by hour band.

False alarm rate, classified into accidental and technical.

Missed check-ins, and the time from each to the first contact attempt.

Near-miss reports per hundred shifts or visits.

Withdrawals.

Escalation tree currency: when it was last tested, and how many numbers were wrong.

Five of the six describe the system. That is deliberate.

What each says when it moves

Response time rising: somebody left the rota, a number changed, or the monitoring arrangement drifted. Check the tree first.

False alarm rate rising: a firmware update, a new device model, or a change in the work. It is rarely the workers.

Missed check-in response lengthening: the receiver is overloaded or the coverage arrangement changed.

Near-miss reports falling: cooperation is being lost. This looks like an improvement on every dashboard and is the most important negative signal here.

Withdrawals at zero: the permission is not believed.

What not to measure

Check-in compliance as a worker performance figure, which turns a safety mechanism into a timekeeping one and destroys the accuracy of both.

Alarm counts per person, which reflect the round rather than the person.

Device carriage as a disciplinary matter, which produces devices carried and not worn — the same nil protection with better numbers.

Anything that makes a worker's use of the system a thing they are judged on.

Who gets what

The programme owner: all six, monthly.

Managers: their team's near misses and any structural findings, not a compliance table.

Senior management: response times against the hazard, and incidents.

The workforce: response times and the false alarm rate.

That last one is unusual and is the point — publishing both says plainly that the system is fallible and that somebody is measuring it.

The annual view

Did response times improve?

Were the structural findings acted on, or only logged?

Did reporting hold or rise?

Is the not-alone list still accurate, and is it being honoured in the schedule?

And one honest statement about what the programme has 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 monthly and read by nobody is an administrative artefact, and the tell is simple: when a number moved last, did anything change?

If the answer is no for a year, the measurement is not the problem — the ownership is.

Producing them without a system

All six come from records most organisations already hold.

Response times from the test log, which is a spreadsheet.

False alarms from the monitoring provider's monthly report, classified by hand — twenty minutes.

Missed check-ins from whoever receives them.

Near misses and withdrawals from the reporting route.

Tree currency from the last test.

None of this needs a dashboard, and a programme that waits for one measures nothing for a year.

The trap of averaging

An average response time across all hours conceals the only figure that matters.

Report by hour band: office hours, evening, overnight.

The overnight figure is typically several times the daytime one and is the one to publish, because the work that generates the risk happens then.

The same applies to sites: an average across a region hides the one location where nothing works.

Setting a baseline

Before any change, record the six figures as they are.

Because without it nothing afterwards can be attributed, and the programme will be judged against a comparison nobody can construct.

This is the step most often skipped in the rush to deploy, and it costs one afternoon at the point where a delay feels least affordable and matters most.

Reporting to a board

One page: response time against the hazard, incidents, and what changed as a result.

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

A report of unbroken success is read once, and the following year's request is harder to make.