The mass-absence playbook — when a third of the floor calls in sick

Operations · ~7 minute read

Norovirus runs through a floor in seventy-two hours. A heatwave closes the trains; a snow day closes the roads; a ballot closes a third of the rota. Mass absence is not a surprise category — it is a known event with known shapes. The playbook exists so that the morning it lands is execution, not deliberation.

The early-warning signals

Almost every mass-absence event telegraphs itself. Norovirus shows up as a Friday absence cluster before it becomes a Monday catastrophe — three sick calls from one team is a signal, not a coincidence. Weather events arrive with days of forecast warning. Industrial action arrives with weeks of statutory notice. School closures, transport strikes, even big sporting fixtures all announce themselves before they bite. The operation that gets ambushed by a mass-absence event is usually an operation that saw the signal and filed it under probably fine.

The playbook therefore starts before the event: a named trigger list (absence cluster above a threshold, amber weather warning, strike notice served), a named person who watches it, and a pre-event checklist — confirm contact details, pre-warn team leaders, line up home-working capacity where it exists, agree the morning call time. Pre-emption is the cheapest lever you will ever pull, and it is the one most often skipped because the event has not happened yet.

The first hour — triage, not heroics

When the morning arrives, the first hour has exactly three jobs. Count: establish actual heads against plan by interval, not a corridor estimate — the difference between 20% down and 35% down changes everything that follows. Reforecast: rebuild the day with the staffing you actually have, so every later decision is made against reality rather than the dead plan. Decide the stance: declare which service levels you are defending and which you are consciously sacrificing today. An operation 30% down cannot hit plan; pretending otherwise just means the sacrifice happens by accident instead of by choice.

The failure pattern in the first hour is the meeting. Eight managers on a call debating whether it is really that bad, while the queue builds and the surviving agents improvise. The playbook replaces the meeting with pre-named decisions: at 15% down we do X, at 25% we do Y, at 35% we declare an incident and stand up the command structure. The thresholds were argued about in calm conditions, precisely so they do not need arguing about now.

Skill prioritisation — the order you wrote down in peacetime

A depleted floor cannot serve every queue, so somebody chooses which queues suffer — the only question is whether the choice is deliberate. The playbook holds a pre-agreed priority order: regulated and safety-critical work first, vulnerable-customer lines protected absolutely, revenue-critical and contractual queues next, and the deferrable work — outbound campaigns, proactive contact, non-urgent back office — named explicitly as the first things to stop. Writing the order down in peacetime is uncomfortable, because it forces stakeholders to accept that their queue is third. That discomfort is the point.

Cross-skilling is the lever that makes the priority order executable. On the day, you collapse the skill matrix towards the protected queues: multi-skilled agents consolidate onto priority work, single-skilled agents on deferrable queues move to whatever they can credibly handle, and the deferrable queues take the published degraded message. None of this can be invented at 9am — the routing profiles for a depleted day should already exist in the platform, tested, named, ready to switch on.

The communications cadence

Three audiences need a rhythm, and each rhythm is pre-named. The floor: team leaders brief at the start and at fixed intervals — what happened, what we are protecting, what we are letting slip, and thanks that sound like you mean them. Agents working a 40-minute queue deserve to know it is a choice, not chaos. Stakeholders: a single situation summary at fixed times — heads, queues, stance, next update — which exists chiefly so that executives do not phone the real-time desk individually for bespoke reassurance. Customers: honest queue messaging and updated digital channels saying what is happening and what to do instead.

The cadence matters more than the content. An update promised for 11:00 that arrives at 11:00 builds trust even when the news is bad; silence breeds the speculation and the side-channel escalations that consume exactly the management attention the day cannot spare. One source of truth, fixed update times, one named voice — everything else is noise.

Recovery — the part everyone skips

Mass-absence events end slowly, not cleanly. Norovirus returns people over a week; a strike may repeat; the backlog from the bad day lands on the merely-bad days that follow. The playbook plans the recovery as deliberately as the crisis: a backlog burn-down with its own priority order, a deliberate decision about whether to use overtime to clear it or let service run warm for a week, and a watchful eye on the agents who carried the worst of it — the thank-you, the early finish, the genuine check-in are operational tools, not soft extras.

Then the loop closes. Within a fortnight, a short honest review: which signals we saw and ignored, which thresholds fired correctly, which pre-named decisions survived contact with reality and which were fantasy. The output is a revised playbook, not a report. An operation that runs this loop gets measurably better at mass absence; an operation that just feels relieved gets the same crisis again with a different virus.

The closing principle

Mass absence is a known event with known shapes, and it always telegraphs itself. The playbook turns the worst morning of the quarter into a sequence of pre-named decisions — who counts, what we protect, what we sacrifice, when we speak — so the floor sees a led operation, not a panicked one.