Green Dashboard, Rising Absenteeism: Early Warning Before Peak Season

The core defect: the thresholds are calibrated to the disaster, not to the drift. A 15% absenteeism threshold means the indicator can only report after the situation is already unmanageable; a fatality threshold means every safety indicator is silent until the worst outcome. Indicators that only fire at the point of failure are not early warning, they are confirmation.

Three changes:

1. Calibrate to baseline, not to catastrophe. Set amber at 1.5x the 12-month rolling baseline and red at 2x, plus a trend rule: three consecutive periods of increase triggers review regardless of level. Under that rule, 4% to 9% over eight weeks would have been amber at week three.

2. Add leading indicators that reflect capacity and supervision, not just outcomes. Candidates: supervisor-to-temporary-worker ratio in the yard; overtime hours per FTE; near-miss reporting rate per 100 FTE (a falling rate is a warning sign, because it usually means under-reporting, not safety); toolbox talk attendance by contractor; turnover among temporary staff; percentage of shifts staffed below minimum crewing; fire service response time to the gate.

3. Build a composite index with weights, reviewed weekly rather than monthly during peak. A weighted index catches the pattern that no single green metric reveals: rising absenteeism plus falling near-miss reports plus contractor disengagement is one signal, not three.

Governance fix: every indicator carries a trend arrow and a time-to-threshold estimate. A dashboard that shows level without direction is how the organisation concluded nothing was broken while eight weeks of drift sat in the data.

The current exercise is compliance theatre: same scenario, attendance counted, nothing changed. Validation should measure what the organisation can do, under conditions it did not choose.

Four design choices:

1. Capability-based objectives. Replace the single tabletop with a three-year cycle targeting distinct capabilities: decision-making under uncertainty, surge staffing, contractor integration, and interface with an external agency. Each exercise states three to five observable objectives and the standard for acceptable performance, agreed before the exercise, not after.

2. Remove the script. Inject events the participants have not seen, including at least one that invalidates the initial assumption (for example, the emergency services arrive later than planned, or the temporary labour agency cannot supply replacements). Run one exercise out of hours and one at shift handover, because that is when real events land.

3. Score with an independent observer against the objectives, and record the time from event to first correct decision and time to correct escalation. Report these as numbers, with a target.

4. Make corrective action closure the deliverable. No exercise is closed until every finding has an owner, a date and a closure test, and the register goes to the executive monthly with closure rate as a reported metric. Set a rule that findings above a threshold cannot be closed by writing a document.

Add one live test per year that is not an exercise at all: a no-notice call cascade or a simulated supplier failure that the operations team must actually resolve within a real working day.

A six-week runway forces triage: stabilise the risk that peak season will expose, and start the redesign in parallel rather than instead.

Weeks 1-2: make the risk visible to people who can act. Present the absenteeism trend, near-miss data and contractor attendance on one page to the executive, with the arithmetic: a third of the yard workforce will be temporary, staffing below minimum crewing is already occurring, and the fire service response time is deteriorating. Do not argue about thresholds; show the drift and the peak-season exposure together. This is the window to change the conversation from nothing is red to these are the conditions we are entering peak with.

Weeks 2-4: fix capacity and supervision. Prioritise the two levers with the fastest effect: supervisory ratio in the yard (cap temporary workers per supervisor, and give supervisors a shift-level stop-work authority they will actually use) and contractor integration (reinstating toolbox talks with contractor attendance recorded, and making attendance a contractual condition). Both are cheap and neither waits on a framework.

Weeks 3-6: prepare the response, not the paperwork. Pre-agree three decision triggers for peak: temporary-worker absence above a set level, crewing below minimum on two consecutive shifts, and any serious near-miss. For each, name who decides and what they can do, including slowing berth allocation. Run one short, unscripted exercise at shift handover, with an independent observer and a written findings list.

In parallel: start the indicator redesign with weekly review during peak, and brief the fire service change into the site emergency plan. Track the peak season as a formal readiness period with a weekly one-page review.