The riot at the opening of a north London branch of Ikea last week may have shown human behaviour at its most unappealing and uncivilised, but not at its most unpredictable.
The crush at the entrance when the store opened its doors at midnight, as shoppers fought to get at the bargains, was precisely the kind of event simulated in a computer model by researchers in Germany and Hungary five years ago. They found that the faster people in a crowd try to pass through a single doorway, the slower they move, and the greater the pressure buildup at the bottleneck.
Dirk Helbing of the Technical University of Dresden and his colleagues were simulating the kind of crowd panic that occurs as people try to flee from a room, for example if a fire has broken out. But their findings apply equally to a crowd eager to get through a door for any reason, such as the offer of a cheap leather sofa.
When an Ikea spokeswoman said that the crowd “behaved like animals”, she was more accurate than she realised. The bargain-hunters behaved, in fact, just like the poor mice that researchers at the University of the Philippines watched in 2003 as they tried to escape through an exit from a flooded chamber. They saw the same clogging effect that Helbing and his colleagues had simulated, which made escape sporadic and inefficient.
Although it is tempting to regard such behaviour as completely irrational, one could argue that in fact it results from an overly focused rationality. The “people” in the computer simulations aren’t knife-wielding maniacs frothing at the mouth, they are automata programmed with a single objective, which is to advance quickly towards the door. They lack the capacity to take a global view of the situation – to see that, if collectively everyone moves a little slower, the door will be used more efficiently.
This sort of dangerous crush at a bottleneck is disturbingly familiar. It is what led to nearly 500 deaths in a fire at the Cocoanut Grove nightclub in Boston in 1942, when the pressure of people trying to get to the doors prevented them from being opened, because they opened inwards. The same effect fatally crushed football fans against riot-control fences at the Hillsborough stadium in 1989, and led to 11 people being trampled to death in the rush for seats at a concert by the Who in Cincinnati in 1979.
But you might think that Ikea would have learned something from the stampede last September at the opening of a store in Jeddah, Saudi Arabia, in which three people were killed. There it was the promise of free vouchers that drew a crowd of 20,000 to the doors, and what followed sounds uncannily like the events in Edmonton. “I am surprised this has happened again after the Saudi incident,” says Helbing.
Yet it wasn’t just inevitable crowd dynamics that led to last week’s injuries. For one thing, the store grossly underestimated the numbers of people the opening would attract, something that infuriated Tottenham MP David Lammy. It lured shoppers with enticing offers, he says, and then “did not put in place the right infrastructure to deal with it”.
Michael Batty, a specialist in urban development and crowd dynamics at University College London, agrees that Ikea got it wrong. “The number of people who went doesn’t surprise me.”
Batty doesn’t think that it was all about discounts, however: there just are not enough stores. “Britain is a grossly under-shopped country compared to the US and France,” he says, “and London is particularly poorly provided for.” He adds that British cities do not have the car infrastructure to cope with big edge-of-city stores.
All the same, there is something uniquely unpleasant about the Ikea experience in London, whether it is at the Brent Park branch in the north or on the Purley Way in the south. “Everybody I know hates it,” says Alan Penn, an urban modeller at UCL, “but everybody goes. I think it’s a form of sado-masochism.” Once you’ve weathered the horrors of getting there, you enter into sensory deprivation in an environment that is manipulative and particularly prone to difficult crowd dynamics.
“They make it a complete nightmare,” says Penn. “You double-back on yourself, you can’t see the shortcuts, you don’t see the outside world. It’s psychologically disruptive, a coercive environment – a kind of brainwashing, really.”
And all cunningly designed for selling. “A phenomenal proportion of purchases at Ikea are made on impulse,” says Penn. “Their whole success is geared to getting people to make impulse purchases.” To do this, the shoppers are forced to follow a tortuous, confusing route through the store. “The signage is not particularly good,” says Batty, “and if you want to go back, it’s not easy.”
Penn and his graduate student Farah Kazim have used a crowd simulation scheme that uses the rules of “space syntax”, developed by Bill Hillier at UCL, to model the movement of people through the Brent Park Ikea store. The model, in which people navigate using their forward-looking lines of sight, reproduces the winding path followed by real shoppers with uncanny accuracy.
Penn is forced almost to admire the fiendish ingenuity of the Ikea scheme. “Shortcuts have to exist for staff and fire regulations,” he explains, “but you don’t see them unless you look back.” Batty agrees: “It’s hard to short-circuit bits of the store.” The problem, however, is that this can cause chaos when the crowd is not passively browsing but is actively seeking out bargains. “It doesn’t help crowd control,” Batty says. For example, Penn explains, if you get “people looking for bargains and squishing their way off the usual route, you get crush points”.
All the same, Penn says, “a crowd disaster is always to do with crowd management having gone wrong” – as at Hillsborough, where it was the opening of an exit gate that caused the crush. Penn says that while such calamities have led to better methods of crowd control in stadiums, “it’s so infrequent in shops that they haven’t learned the lessons”.
Batty adds: “Ikea could learn something from these models. At the moment, they design for selling rather than for crowd control.”
· Philip Ball is the author of Critical Mass: how one thing leads to another (Heinemann)