A Working Arrangement
What this looks like when it is working, assembled from everything here, as a description to measure yours against.
Not a model. A description of an operation where staffing is settled, so that you can see which parts yours is missing.
The people issue in “A Working Arrangement” should be settled before workforce software is configured. For teams considering remote employee monitoring software in relation to remote employee monitoring software, the rollout should state the purpose, explain what is recorded, limit manager access and give employees a practical correction route.
What is known and written down
Loaded hourly cost, recalculated annually.
For an independent reference related to “A Working Arrangement”, consult the ILO guide to balanced working-time arrangements; it provides a useful external check on scheduling, working-time and workforce-planning assumptions.
The demand curve by hour of day and day of week.
The fixed minimum and the deferrable task list.
The conversion factor, measured.
Forecast error and its direction.
The cost of a short hour and a surplus hour, roughly.
Six things, on four spreadsheet tabs.
How the week is built
Forecast written down before the week: last year, plus trend, plus the events calendar.
Hours derived from the forecast by the conversion factor, shaped to the curve rather than to a daily total.
Checked against the skills matrix: capabilities present, not just headcount.
Published two weeks ahead and honoured.
What happens when it goes wrong
The error direction was decided by period in advance, so the correction is already chosen.
A real call list of people who want extra hours.
An agreed minimum payment if a shift is cancelled.
And a stated point at which you stop calling and run short deliberately.
What the reporting contains
Demand forecast and actual.
Hours and cost, variance split by cause.
One service measure.
The ordinary range marked, so noise is not explained.
How a bad week is handled
The six causes worked in order, ten minutes.
Four sentences sent on the day the figures land, unprompted: what moved, why, what avoiding it would have cost, what changes.
And the occasional honest "the schedule was loose", which is why the other explanations are believed.
What the people see
A published horizon that holds.
A stated rule for extra hours and unpopular shifts.
Preferences collected and used.
A standing question about what is coming up, with visible responses.
And nobody carrying the operation alone, because the single points were trained out.
What it cost to get here
Two or three days of analysis, spread over months.
An hour a week of forecast recording.
One day a year of review.
No purchase required, though a system may help later and will be configured properly when it arrives.
The test
Could you say, in a figure, what being one person short on Saturday costs you?
Do you know how wrong your forecast usually is?
Can everybody take a fortnight off?
An operation answering those three has the rest.