The 15-person rule: why a champion network catches what a 500-person worksite can’t

It usually comes out the same way in the room. You run a large multi-site operation. The wellbeing program reaches the whole headcount — that is what the reporting says, that is what the EAP utilisation deck shows. And yet the people who are most at risk are the ones the program never seems to touch. Worse: you keep being surprised by incidents that, in hindsight, someone standing three metres away could have seen coming. The information existed. It just never reached anyone who could act on it.

That is not a coverage problem. It is a structural one. And it is worth being precise about why, because the usual fix — more reach, more comms, more sign-ups — makes the number on the dashboard go up without changing the thing that actually failed.

Why does "whole-headcount coverage" miss the people most at risk?

Coverage measures access. It tells you that a service exists and that, in theory, anyone can reach it. It says nothing about whether the early signals of psychosocial harm — the withdrawal, the uncharacteristic mistakes, the second shift that becomes a pattern — are being noticed by anyone close enough to read them.

Those signals are not visible from head office. They are not visible from an annual survey. They are visible to the people who work alongside the person every day, who know what "normal" looks like for that individual and can therefore see the deviation. The leading indicators live at the crew. Your detection capability, in most organisations, lives several layers above it.

The default answer to "who notices?" is the supervisor. That answer is structurally weak. Research into the Australian construction workforce found that workers identify their supervisors as among the least trusted resources for raising a mental-health concern, and that peer-led approaches consistently outperform supervisor-led ones (Gullestrup et al., Health Promotion International, 2023). The person with the org-chart responsibility to notice is frequently the last person a worker will open up to — because that same person controls their shifts, their assessment, their progression. The reporting line and the trust line run in opposite directions.

What does Dunbar's number have to do with people-risk?

There is a hard ceiling on the number of stable relationships any one person can actually hold. The figure is roughly 150 — Dunbar's number. Inside that, there is a much tighter band of perhaps 15 people you genuinely know well enough to read.

This is the part that matters operationally. You cannot know 500 people. No leader can. The moment your worksite crosses 150, you are past the limit at which any single point of oversight can hold the relationships needed to see what is happening. The human limit is 150. The risk begins at 151.

So the question "how do we make our wellbeing program reach 500 people?" is the wrong question. One program, broadcast from the centre, cannot read 500 people no matter how well funded it is. The right question is: who already knows the 15? Because someone does. On every crew, on every shift, there is someone the others actually talk to. That person already has the relationships. What they do not have is the trained eye, the language, and the structured route to escalate what they see.

What is the difference between a poster and an infrastructure?

A poster — and a one-off resilience training day, and an awareness month — pushes information outward and hopes it lands. It is centred on the assumption that if people just knew the signs, the system would work. It treats the gap as a knowledge problem.

The gap is not a knowledge problem. It is a structural one. The people who can see the early signals are not connected to the people who can act on them, and there is no operating model that moves a quiet observation from a crew member into a controlled response before it becomes an incident.

A champion network is that operating model. It takes the trusted person who already knows the 15 and equips them — deliberately, with structure, not as an "accidental counsellor" left to improvise — to recognise what they are seeing, hold a first conversation, and route the person to the right help at the right level. It sits upstream of the EAP. The EAP is reactive support that activates after harm has surfaced; the champion network is the layer that catches the signal earlier, in the gap where nothing structured currently exists.

This matters doubly in a workforce where the people most at risk are often the hardest for a centralised program to reach. Construction is roughly 88% male against a workforce average near 53% (Gullestrup et al., 2023), which compounds known help-seeking barriers. And an estimated 27% of the Australian construction workforce — about 255,000 workers — was born overseas (Maheen et al., The Lancet Regional Health – Western Pacific, 2024), which means a program delivered in one register, through one channel, in one language, is structurally blind to a quarter of the people it claims to cover. A champion on the crew speaks the language of the crew — literally and culturally — in a way no centralised broadcast can.

The peer-led model is not theoretical in this sector. MATES in Construction runs a champion-based, peer-led approach across construction, mining, energy and manufacturing (MATES, 2024) — built on the same observation: that the worker next to you, trained and connected to a route, sees and is trusted in a way the org chart never will be.

How many champions do you actually need?

As a working ratio, about one trained champion per 30 staff — adjusted for the number of sites, the geography, and how dispersed and transient the workforce is. A single 500-person worksite and a 500-person operation spread across twelve sites are not the same problem, and the network is sized to the second number, not the first.

The point of the ratio is not headcount theatre. It is to guarantee that no crew sits outside the trained eye — that every band of roughly 15 to 30 people has someone within it who can see the deviation and knows what to do with it. Below that density, you reopen the gap. Above it, you are buying redundancy you do not need.

Why is this an operations and WHS matter, not an HR one?

Psychosocial hazards now sit inside your WHS duty. They belong on the risk register, they are subject to the hierarchy of controls, and they are squarely within an officer's due diligence obligation to do what is reasonably practicable. A champion network is a control — a structured, documented, upstream control that you can point to. A poster is not a control, and "we have an EAP" is evidence of a reactive support arrangement, not of prevention.

This is why the buyer is operations, not wellbeing-as-a-nice-to-have. The exposure — incident, claim, regulator, insurer, the downstream productivity and turnover cost of harm that surfaced too late — sits with the operator. Running a champion network as part of a structured psychosocial safety framework does two things at once that look contradictory and are not: it reduces the load on the EAP by catching people before they escalate into crisis, and it increases appropriate EAP uptake, because people get routed to the right help earlier instead of falling through. Fewer crises, better-aimed support. Both.

The 15-person rule is not a wellbeing slogan. It is an honest statement about the limits of human attention, applied to a risk you are already legally on the hook to control. You cannot know 500 people. You do not have to. You have to make sure someone knows their 15 — and that what they see has somewhere to go.