The Station nightclub: 90 seconds, four exits, one door

There were four exits. The crowd went for one. On 20 February 2003, when pyrotechnics ignited soundproofing foam at the Station nightclub in Rhode Island, the majority of people inside made for the front door they had walked in through, and the crush there is where most of the hundred deaths happened.

The foam was there to solve a noise complaint

The polyurethane foam on the walls had been fitted to reduce noise for neighbours. It was not rated for the purpose, it was never intended to be a wall lining, and it turned a small building into an untenable one in roughly ninety seconds. The NIST investigation reconstructed it in detail, and the timeline is the most sobering document in modern fire engineering.

But the material is not the transferable lesson. Plenty of buildings have had bad linings and evacuated successfully. What killed people here was the geometry of a crowd and its habits.

Occupants are not evacuees, they are people going back the way they came

One hundred people died and roughly 230 were injured. The timeline reconstructed afterwards is the part that stays with fire engineers: the pyrotechnics fired at about 23:07, the first alarm sounded some thirty seconds later, and the building was no longer survivable inside about ninety seconds of ignition. Video shot by a news crew who happened to be filming that night captured the whole of it, which is why this fire is one of the most precisely documented in history. Survivors were treated for burns and inhalation injuries across half a dozen hospitals, and many needed years of reconstructive surgery.

Human behaviour in fire is consistent and well documented: people move towards the familiar, they underestimate the seriousness for the first crucial seconds, and they follow whoever seems to know what they are doing. Exits they have never used feel like staff areas, and a door they have not been through is barely a door at all.

That is why crowd managers became mandatory in American assembly occupancies after this fire, and why NFPA 101 extended sprinkler requirements to existing nightclubs rather than only new ones. It is also why your evacuation strategy cannot stop at exit widths.

The version of this in your building

The bill took seven years to settle and came to 176 million dollars, shared among more than three hundred claimants, and paid by a long list of defendants who had each contributed a small piece of the failure: the band's management, the club's owners, the foam supplier and installer, the brewer whose promotion filled the room, the broadcaster whose crew was filming, and the town whose inspector had passed the premises. Nobody involved thought they were running a fire risk. Every one of them paid.

That distribution is the lesson for a landlord. When a venue in your building fails, the claim does not stop at the tenant who fitted the foam. It reaches whoever approved the fit-out, whoever held the fire risk assessment for the common parts, and whoever could have asked for the certificate and did not. Your name appears on that list because of what you could have required, not because of what you did.

Your occupiers arrive through one entrance every day, five days a week, for years. That is the route their feet know. The secondary staircase they have never used, the final exit through the service corridor, the door at the back of the retail unit — those exist on your drawing and not in their heads.

Evacuation drills are the only mechanism that changes this, which is why a drill that releases everyone through the main entrance is worse than no drill at all: it rehearses exactly the behaviour you need to break.

What the foam did, and how fast

The material on those walls was flexible polyurethane foam, chosen as sound treatment and entirely unsuitable as a wall lining. Its surface area is enormous relative to its mass, so it ignites readily and releases heat extremely quickly, and its combustion products include hydrogen cyanide alongside carbon monoxide. NIST's reconstruction put the room beyond tenable conditions in roughly ninety seconds from ignition, with flame spreading across the ceiling and smoke descending to head height well before most people had accepted that anything was seriously wrong.

NIST also modelled the counterfactual, which is why this investigation matters. With sprinklers, the simulations showed conditions would have remained survivable for long enough for the room to clear. The building had none, because it fell under the threshold at which existing premises were required to have them.

Crowd managers, capacity and the maths of getting out

The code changes that followed are unusually concrete. Assembly occupancies above a threshold occupant load require sprinklers in new and existing premises. Trained crowd managers are required, at a ratio of roughly one per 250 occupants above the first 250, and their job is exactly what nobody did on the night: direct people to exits they would not choose for themselves. Indoor pyrotechnics require a permit, separation distances and a responsible person.

The crowd manager provision is the one with the widest application. Every exit calculation assumes the population distributes across the available exits, and that assumption is false in every building unless somebody makes it true on the night. A competent person standing at a secondary exit is not a formality; it is the mechanism by which your arithmetic becomes real.

Your 60-second self-check

• When you last ran a drill, did any group leave by a route they do not use daily?
• Are your secondary escape routes signed, lit and clear enough that a visitor would believe they are a way out?
• Does anyone have the role of standing at an alternative exit during an evacuation and sending people through it?

An exit that is theoretically available and practically unknown does not count towards your capacity. The free 5-minute Director's Risk Check maps which duties and deadlines apply to your portfolio, each with its official source, so you can see where the exposure sits before someone else finds it for you. Run it at the Director's Risk Check, or send it to whoever holds the files. You can only count the exits your occupiers would actually use, and only a drill tells you which those are.

Herman's verdict

People leave by the door they came in through. Every alternative route in your strategy is an assumption until somebody is standing at it, on the night, telling people to use it.

— Herman