How many people you actually have CAPACITY

Two numbers, and almost everybody quotes the wrong one

The establishment is how many posts exist. The strength is how many of them are filled. Neither is how many people will be at work on a Tuesday.

The difference is not a rounding error. Take off annual leave, training, sickness, maternity and paternity, restricted duties, people seconded elsewhere and whatever your trade calls being abstracted, and a nominal thirty is very often a real nineteen. That is the number the rota has to be built from, and it is the number nobody has to hand.

The arithmetic is trivial. Almost nobody does it, because the abstractions live in five different places and each one is somebody else's spreadsheet: leave in one system, training in another, sickness with HR, secondments in a memo, restricted duties in the head of a supervisor. So this puts them in one column, takes them off in order, and shows the honest number next to the headline one.

Why quoting the establishment causes real harm

It is not a presentational problem. A plan built on thirty when nineteen turn up produces a rota that cannot be worked, and what happens next is predictable: the gap is covered by overtime, by cancelling rest days, or by running short and hoping. All three are paid for, and two of them are paid for in sickness, which makes the number worse the following month.

It also makes the case for more staff impossible to argue. “We have thirty and cannot cope” invites the answer that thirty ought to be enough. “We have thirty posts, twenty-six filled, and after leave, training and restricted duties we field nineteen on an average Tuesday, against a minimum safe number of twenty-two” is an argument with a number in it, and it is the same claim.

Being honest about the deductions

Use the actual rates, not the policy ones. Annual leave is a known entitlement and it is the easy one. Sickness should come from your own last twelve months rather than a national average, because it varies enormously between teams and the local figure is the one you will live with. Training days are known in advance and are almost always understated, because the mandatory refreshers get forgotten until the week they fall due.

Count the partial abstractions properly. Somebody on restricted duties is not absent; they are available for some of the work and not the rest, and recording them as present overstates the rota while recording them as absent is unfair and also wrong. Half a person is a better number than either.

Recalculate when the inputs change rather than annually. A team that loses two people in March has a different capacity from April, and a plan still quoting the January figure in June is describing an organisation that no longer exists.

Nothing you type here is uploaded, which matters because the inputs are sickness and restricted duties attached to a small team: personal data about identifiable people, even when the page shows only totals. It stays in this browser. Nearby: ROSTER builds the rota from the number this gives you, LEAVE shows the days you go under, ONCALL works out who is on, and SLA answers when a response is actually due in open hours.