The Cost That Never Reaches Finance

Agency invoices arrive monthly. Overtime arrives in payroll. Both are reviewed, both are too high, and neither carries any information about what caused it.

The record in “The Cost That Never Reaches Finance” becomes more useful when the original time entry, later correction and reviewer decision remain distinguishable. If more information supports workforce analytics software, administrators should test exports, amendments, permissions and retention before launch so a dashboard does not replace the underlying evidence.

So the response is a procurement response: renegotiate the agency rate, cap overtime, question the volume. These produce modest savings and leave the cause untouched, because the cause is upstream in a system finance does not see.

For an independent benchmark relevant to “The Cost That Never Reaches Finance”, consult the Grants.gov policy resources. Use it to test scheduling, working-time limits, attendance records, employee rights and exception handling against the real operation rather than treating a software report as self-explanatory evidence.

The join

Cover spend has a date, a site, and usually a shift. The morning log has the same three. Joining them attributes spend to cause, and that is the whole exercise.

Where the morning log does not exist, the absence record is a weaker substitute: absences by date and team, matched against agency bookings for the same date and team. It is imprecise and it is enough to show the shape.

The output is a figure most operations have never seen: cover spend attributable to absence, separated from cover spend attributable to vacancy, and both separated from planned cover for peaks. Three numbers with completely different remedies.

Why they get conflated

Agency spend is reported as one line. Within it, covering a vacancy for three months while recruiting, covering a known seasonal peak, and covering yesterday's sickness are indistinguishable.

They need different responses. The vacancy is a recruitment problem. The peak is a planning problem. The absence is everything in this collection. Treating them as one produces the standard response — reduce agency usage — which is not a remedy for any of the three.

What else should be in the figure

Overtime attributable to cover, separated from overtime for planned work.

Bank hours.

Management time spent covering, which requires the record described elsewhere and is otherwise free and invisible.

The hour itself, estimated.

The first two come from payroll with a reason code. The third requires somebody to have written it down. The fourth is an estimate and should be labelled as one.

Giving it an owner

The reason this join does not happen is that nobody owns both sides. Operations owns the mornings and finance owns the spend, and the number that would connect them is in neither remit.

Making it somebody's job — a quarterly page, produced by whoever owns the morning log, reviewed with finance — is the whole intervention. It is an afternoon a quarter.

What it changes

The conversation moves from "agency spend is too high" to "we spent this much covering absence, of which this much came from six specific capability gaps and three brittle shift patterns, and here is what fixing each would cost".

That is a business case rather than a complaint, and it is approvable. Every intervention in this collection — a second trained person, a float post, a working call-out list, a change to a turnaround — is cheap against the quarterly figure and unjustifiable without it.

The number to lead with

Cover spend attributable to absence, per quarter, as a proportion of pay cost for the same population.

One number, one denominator, trended. It is the single most useful figure this subject produces and in most organisations nobody has ever calculated it.

Making it survive a change of owner

A quarterly figure produced by one enthusiastic person stops the month they move on, and the organisation reverts to managing agency spend as a procurement line.

Write the query down — literally, the steps and the sources — and put the page on a standing agenda somewhere. A number that appears on an agenda has to be produced by somebody; one that depends on goodwill does not. It is an unglamorous piece of institutionalising and it is the difference between a one-off analysis and a measure.

The three questions finance will ask

Whether the attribution is reliable, which it is not perfectly and does not need to be — a figure accurate to within a fifth is enough to change a decision that is currently being made on no figure at all.

Whether the spend would disappear if the fix were made, which it would not entirely, and the honest answer is a proportion with a reason. And whether this is really a staffing argument in disguise, which it sometimes is, and saying so directly is better than being caught at it. Preparing those three answers before the meeting is what distinguishes a figure that gets acted on from one that gets queried into irrelevance.