Forget the new year’s resolution: September, not January, is the time for new starts. College freshers are preparing to leave home, graduates are ironing shirts and blouses and dressing up for their first day in the office. Recruiters and admissions tutors are hoping they made the right choices.
So how do we select the best people for a course or a job? It seems like a sensible question, yet it contains a trap. In selecting the best person we might set a test — in a restaurant kitchen we might ask them to whip up some meals; in a software company we might set some coding problems. And then the trap is sprung.
By setting the same task for every applicant we recruit people who are carbon copies of each other. They will have the same skills and think in the same way. Allowing recruiters some subjective discretion might loosen this trap a little, but it might equally make it worse: we all tend to see merit in applicants who look, speak, and dress much like we do. Opposites do not attract, especially when it comes to corporate hiring.
This is unfair, of course. But it is also — for many but not all tasks — very unwise. Scott Page, a complexity scientist and author of a new book, The Diversity Bonus (UK) (US), invites us to think of people as possessing a kind of cognitive toolbox. The tools might be anything from fluent Mandarin to knowing how to dress a turkey to a command of Excel keyboard shortcuts. If the range of skills — the size of the toolkit — matters, then a diverse team will boast more cognitive skills than a homogenous team, even one full of top performers.
The logic of this toolbox model is obvious in certain contexts: any good heist movie will have the bruiser, the charmer, the hacker, the explosives expert, the strategist and the cat burglar. It is clear why such a diverse range of skills is needed and it is obvious that no single test could recruit such a team: it has to be constructed with diversity in mind from the start.
But within a corporate environment the same logic often tends to be forgotten. Everyone has been recruited using the same cookie-cutter template; everyone is proficient at a similar set of tasks, and the range of thinking skills suffers. The IMF is full of economists. Congress is full of lawyers. Football management is full of ex-footballers. If someone does happen to have hidden talents that will be by accident, not by design.
This homogeneity may not be disastrous. If you want to recruit 10 truck drivers you probably just need the 10 safest, most reliable drivers you can find, because the drivers will be working as individuals, not sparking off each other. But in any situation where a range of problems have to be solved together as a team, diversity can help.
Scott Page’s model of diversity — less a glorious rainbow of superficial attributes, more a toolkit crammed with different skills and perspectives — is a powerful way to appreciate the problem with homogeneity. If recruiters keep looking for the same skillset then an organisation risks, in the words of philosopher-queen Alanis Morissette, having 10,000 spoons when all it needs is a knife.
The standard model of graduate recruitment is almost helpless when faced with this problem. Yes, one can have diversity coaching, checking that certain demographic groups aren’t being discriminated against. But when one candidate at a time is being recruited, it is hard to do much about diversity because diversity is not a property of individuals, it is a property of groups.
So how to solve the problem? Rory Sutherland of Ogilvy, an adman with a keen interest in behavioural science, has suggested recruiting people in groups. If an organisation recruits five people at a time then a couple of vacancies can be reserved for wild-cards — people who don’t fit the mould but have interesting talents. But the definition of “interesting” is itself a tricky one, and not every organisation has the luxury of recruiting in bulk.
What makes matters worse is that we often do not appreciate the value of diversity when we see it. One study of problem-solving (by Katherine Philips, Katie Liljenquist and Margaret Neale) found that groups containing an outsider were far more proficient at solving murder-mystery puzzles than groups made up entirely of friends.
The striking thing about this study, though, was that the successful groups with an outsider didn’t realise they were being successful, while the cosy underperforming groups of friends were complacent, not realising how badly they were doing. Having the outsider around helps us solve problems, but don’t expect us to be grateful, or even to notice anything other than social discomfort.
The hard truth is that to find new solutions to old problems we must often work with people we don’t really understand. I won’t pretend this is easy, but I cannot wait to pull off a heist with Scott Page, Rory Sutherland and Alanis Morrissette.
Written for and first published in the Financial Times on 8 September 2017.