Starters, Leavers and Short Service

An averaged limit assumes a person has been present for the length of the reference period. New starters have not, and the regime has to say something about what happens in the meantime.

The workflow in “Starters, Leavers and Short Service” becomes more reliable when scheduled hours, actual time and later corrections can be distinguished. For teams exploring mouse jiggler detection, review the platform here can add operational time and project context, provided data collection is proportionate, permissions are limited and every important exception receives human review.

Most regimes handle it by averaging over the period actually worked, which is sensible and has a consequence organisations rarely anticipate: a new starter's average is far more volatile than anybody else's, because it is computed over a handful of weeks rather than seventeen.

For an independent reference relevant to “Starters, Leavers and Short Service”, consult the GOV.UK working-hours 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 a short history is volatile

With five weeks of data, one sixty-hour week moves the average by more than two hours. With seventeen weeks, the same week moves it by less than a third of that.

The practical effect is that new starters cross limits quickly and recover from them slowly, which is the opposite of what the rota planner expects. Somebody three weeks into the job who covers a heavy fortnight can be over the limit before anybody has thought about them, and it will take months of ordinary weeks to bring the figure back down.

The group most likely to be in this position

New starters are frequently the people asked to work extra: they are keen, they want the money, they have not yet learned to decline, and they are often recruited precisely because the department is short.

So the group with the most volatile figure is the group most likely to be given hours, and the group least likely to raise it. That combination is worth a specific check rather than relying on the general report, because the general report tends to be read by exception and a new starter's name means nothing to whoever reads it.

What the system usually does

One of three things, and it matters which. It computes over weeks actually worked, which is usually correct. It computes over the full period with missing weeks as zero, which produces an artificially low figure and large apparent headroom. Or it excludes people with insufficient history from the report entirely, which is the worst outcome because they disappear.

The third is common and almost always unintentional: a report filter that requires a complete period in order to produce a valid average.

Leavers

The mirror image, and it matters for a narrower reason: the final period is the one where a breach cannot be corrected by future light weeks.

Somebody working a notice period is often clearing work, handing over and covering gaps, and their last six weeks are frequently their heaviest. Nothing after that will average it down. Checking the figure at the point notice is given, rather than at the end, is the only moment at which anything can be done.

Transfers and internal moves

A person moving from one site or department to another is not a leaver and is frequently treated as one by the data. Their history stops at the old record and restarts at the new one.

That produces exactly the short-history volatility described above, for somebody who has been with the organisation for years, and it is entirely an artefact of how the records are keyed. Where this happens, the fix is a report that follows the person rather than the assignment, and it is the same fix needed for the multi-site problem described later in this collection.

Agency placements that become permanent

Common in warehousing, care and manufacturing: twelve weeks through an agency, then a direct contract. The hours before the transfer were worked by the same body in the same building.

Whether they count towards the average under the new contract depends on the regime and on how the arrangement is structured, and it is a question worth asking once rather than assuming. What is not in doubt is that the person is as tired as their total hours make them, whoever was paying.

What to check

Three queries. Everybody with less than a full reference period of history, with their average computed over the weeks they have. Everybody serving notice, with their figure to date. And everybody who transferred internally in the last six months, with their combined history across both records.

None of those appears in a standard report. All three are the same data with a different filter, and each identifies a group that the ordinary monitoring is structurally blind to.

The seasonal intake

An organisation taking on fifty people for a peak has fifty volatile figures at once, in the weeks when demand is highest and everybody is being asked to do more.

Treat the cohort as a group rather than as individuals: one check across all of them, weekly, for the first eight weeks. It is the same query with a date filter, it catches the problem while the peak is still running, and it avoids the usual outcome where the breaches are discovered in January along with everything else from the busy period.