What Not to Do

Each of these is common, each is understandable, and each costs more than it saves.

Scheduling to the forecast exactly

Why not: every variance becomes a cut or a call-in, and in several jurisdictions each cut costs a premium.

Instead: a buffer sized from measured forecast error.

Using availability breadth to allocate hours

Why not: it penalises people with caring responsibilities and correlates with protected characteristics, which makes an allocation practice a discrimination question.

Instead: a stated allocation rule, applied consistently, with the distribution published.

Traditional on-call

Why not: it takes a person's whole day for nothing, and most fair workweek jurisdictions restrict or prohibit it.

Instead: a voluntary standby list.

Copying last week forward

Why not: it perpetuates whatever unfairness was already there and hides it behind continuity.

Instead: a deliberate pattern, and a quarterly look at the slot distribution.

Cutting shifts to hit a labour percentage

Why not: the saving is visible and the lost sales are not, so the metric rewards a decision that loses money.

Instead: report service measures and turnover alongside the percentage.

Making the sick person find cover

Why not: they are unwell, and they will ask whoever is easiest rather than whoever is right.

Instead: a stated order of recourse, owned by the manager.

Refusing swaps to enforce the original rota

Why not: it produces informal swaps and a record that is fiction.

Instead: confirm on capability, rest and hours — not on preference.

Treating warnings as constraints

Why not: a warning at eight in the evening gets clicked through, and the rest breach is published.

Instead: hard constraints for rest, minor limits and capabilities.

Announcing a change before fixing anything

Why not: it produces defensive behaviour and contaminates the baseline you were about to measure.

Instead: fix one visible thing, then say what you fixed.

The pattern behind them

Most transfer uncertainty or cost from the business to the people, invisibly and in small amounts.

They are individually minor and cumulatively the reason these sectors have the turnover they do.

The test for any practice

One question that catches most of these before they start.

Would you write this rule down and show it to the people it applies to?

"Awkward shifts go to whoever complains least" fails immediately.

"Hours go to whoever is most available" fails once stated.

Anything that only works while unstated is the thing to change, and stating it is how you find out.

Reproduce the workflow

For another way to make this requirement testable, review inspect this service connection. Treat it as a starting point, then reproduce the case with real roles, sites and exceptions.

Independent reference

For a thematic point of reference, see the International Labour Organization. Its current material provides useful context beyond product documentation.