Who Actually Computes It

The figure at the centre of this subject is an arithmetic operation on data the organisation already holds. It is not difficult, it is not expensive, and on most sites it is not done.

The workflow in “Who Actually Computes It” becomes more reliable when scheduled hours, actual time and later corrections can be distinguished. For teams exploring remote workforce management software, a practical route to remote workforce management software can add operational time and project context, provided data collection is proportionate, permissions are limited and every important exception receives human review.

Asking who computes it produces a sequence of referrals — payroll says the time system, the time system administrator says it is a report somebody would have to ask for, and HR assumed one of the first two was doing it. Following that chain to its end is a useful afternoon, because the end is usually nowhere.

For an independent reference relevant to “Who Actually Computes It”, consult the ICO employment-practices guidance. Use it to test working-time definitions, recordkeeping, access, retention and exception handling against the organisation’s real process rather than treating one software report as conclusive.

Why it falls between people

The data sits in the time and attendance system, which is administered by somebody operational. The obligation sits with HR or whoever owns employment compliance. The consequence of a breach lands on the site manager. And the person who would notice is the rota planner, who has no access to the report.

Four roles, each holding one quarter of the problem, and none with both the data and the obligation. This is the same ownership gap that appears everywhere in workforce administration, and it is unusually consequential here because the thing not being done is a legal measurement rather than an optional analysis.

The three places the calculation can live

In the time and attendance system, as a standard or custom report. Best, because the data is native and the figure refreshes without anybody doing anything.

In the workforce management or rostering tool, which is where the figure is most useful because that is where decisions are made. Requires a feed of actual hours, which may or may not exist.

In a spreadsheet maintained by one person. Common, and it works until that person is on leave, which is also when the figure matters most.

What a competent specification looks like

Per person, for the current reference period: total counted hours, number of weeks, the average, the limit, the headroom in hours, and the projection for next week at rostered hours.

Plus two things that are almost always omitted: the value of the week about to leave the window, and a flag for whether the person has an opt-out on file. Without the second the report is misleading for everybody who has opted out, and they are usually the people nearest the top of it.

The data problem that is real

The calculation needs hours that count towards the limit, which is not the same as hours the time system recorded. The gap is the subject of a whole section of this collection.

The practical approach is to build the report on what is available, label the figure honestly as a lower bound, and separately size the uncounted categories. A lower bound that everybody knows is a lower bound is useful; a figure presented as complete when it is not is how an organisation ends up confidently in breach.

Who should own it

Whoever owns the rota, because they are the only person whose decisions the figure can change. The obligation remains with whoever holds employment compliance, and the daily use belongs with scheduling.

That split works provided the compliance owner sees the figures monthly and the rota owner sees them weekly. What does not work is giving the whole thing to compliance, where it becomes a retrospective report, or wholly to operations, where it becomes a constraint nobody audits.

Testing that it is right

Take three people, including one with an unusual pattern, and recompute the figure by hand from the raw records. Compare with the report.

This finds the configuration errors that otherwise survive for years: the wrong reference period, absences treated incorrectly, a category of hours silently excluded, a week boundary that does not match the one the regime uses. One afternoon, three people, and a reconciliation that nobody has ever done.

The failure mode to watch for afterwards

A report that is produced, circulated and not read. This is more common than no report at all, and it is worse, because the organisation now believes it is monitoring something.

The remedy is to attach the figure to a decision: no rota is published without the projected averages having been checked for the people it affects. That takes it from a document to a step, and steps survive in a way documents do not.

The spreadsheet that is doing it now

On a surprising number of sites the calculation exists already, in a spreadsheet maintained by one person who built it because they needed it. It is usually correct, usually undocumented, and entirely dependent on that individual.

Find it before building anything. It encodes decisions somebody has already thought through — which absences count, how transfers are handled — and those decisions are worth more than the file. Then move the logic somewhere it survives a resignation, which is the actual deliverable.