How Metaview is levelling the playing field in job interviews

We’ve all messed up job interviews, but what about the interviewers, can they mess up too? Andy Martin speaks to the founders of Metaview about using data to help people find their perfect job

Metaview has assembled a database of tens of thousands of interviews and can tell you exactly what is going wrong  (Getty/iStock)

How many times do you walk out of an interview thinking, I blew it? Rather like performing badly in an exam, everyone at one time or another (in my own case more than once) has given a disastrous interview. And you regret what might have been. But consider this: what about the people on the other side of the table, the ones who are doing the interviewing? What if they blow it too? Then you end up with a bad appointment and the company misses out on an opportunity to build its team. Now, fortunately, no one has to blow it any more – not if Metaview has anything to do with it, thanks to its co-founders, Siadhal Magos and Shahriar Tajbakhsh.

The whole point of Metaview (https://www.metaview.app/), founded in 2018, is to put an end to all those screw-ups. Tajbakhsh himself, now CTO, reckons he knows all about how to fail an interview. He left Iran for England aged 16 and ended up studying computer science at UCL. He liked the look of Palantir, who specialise in big data, and duly applied for a job. And was duly rejected. But he didn’t give up – he kept on applying, year after year, until finally, on the fifth time of trying, they finally surrendered and gave him the job. “I was just incompetent,” he says now. “But I managed to slip through. The hack is to work hard.”

Magos, who is Irish-Hungarian and CEO, overlapped with Tajbakhsh at UCL where he was studying history. He went to work for Uber, which grew from a staff of 7,000 to 17,000 in two years, and he found himself conducting a lot of interviews (as well as delighting all Uber drivers by coming up with a system called “Instantpay”). Tajbakhsh, similarly, at Palantir (a mere 2,500), was part of a committee making the final decision on all those would-be recruits. They decided to put their heads together and pool their knowledge and create Metaview (an elite crew of 11). They have now assembled a grand database of tens of thousands of interviews and can tell you exactly what is going wrong and teach you how to get it right.

They have some fairly shocking statistics at their fingertips. For example, if you’re a woman interviewee, do you ever have the feeling you’ve been undervalued in comparison to your male competitors? Are the interviewers being sexist? It turns out you were right all along. Women are typically given 12 per cent less time to speak in interviews than men. “It’s been like this for decades,” says Magos. “Before it was imperceptible. The interview was a black box. Now we have the data.” Metaview is levelling the playing field, to make sure that everyone gets a fair crack of the whip, irrespective of gender, race, or hairstyle.

Magos says that the key to the process is “rigour and consistency”. In the long term, the data will show up bias. Metaview also has a system of “automatic notes”, so that no one has to waste time scribbling on a pad or tapping on a tablet and can concentrate on learning about the candidate. Some interviewers can be too “dominant” and speak for longer than the people they are interviewing. Bad move. After all the hopeful no-hopers have been shaken out, the interview process will generally boil down to a handful of high-calibre candidates only. “We are talking about not just the top 1 per cent,” says Tajbakhsh. “It’s more like the top 0.1 per cent.” You have to be nice to them.

They don’t have to go and work for Netflix, they can always go down the road and work for Google instead. If you’re going to behave like a Grand Inquisitor they won’t even want to come and work for you. A good interview involves a “mutually beneficial exchange” in which each side gets to know more about the other. “We don’t do filtering,” says Magos. “We do matchmaking.”

Interviewers can’t just ask a set of pre-prepared questions. They have to be able to engage with the interviewee

The funny thing about interviewing the Metaview guys is that they were actually interviewing me. And they’re good at it. They wanted to know not just what I did at The Independent (this, for example), but also whether I had made a significant contribution to increasing subscribers (I know of one guy, definitely). The important point they stressed was that the interviewer needs to zero in on “outcomes” and “impact”. Magos adds, “It’s not so much the what, it’s more the why: why do you do what you do? What inspires you? Is there a legitimate passion for specifically what we are trying to achieve?”

Whether for good or ill, you don’t walk into a room any more, sit down, and get grilled by a committee. Most interviews are now remote, conducted in virtual rooms online. I still think you had better comb your hair and look presentable. But given that so much interviewing now takes place via machines, couldn’t we just cut out the middle man entirely and let the computer decide?

“In an ideal world,” says Magos, “you might have an AI that could pass the Turing Test and would be indistinguishable from a human interviewer. But the reality is we are not  there yet.” A machine cannot replicate the “final interview” (each individual has to go through an average of four to six interviews, sometimes as many as a dozen). Magos stresses that most of the companies that Metaview work with (over 300 of them in several countries, such as Bulb, Careem, AngelList, Sendwave) are on the creative side, whether in design or software. And here the human factor is still crucial. It’s not just a matter of pure data and binary decisions, people have to know if they are going to click with the potential newby.

“Do you share the same values? That is the crucial question,” says Magos. Interviewers can’t just ask a set of pre-prepared questions. They have to be able to engage with the interviewee and ask follow-up questions to their answers, burrowing down deeper into their experience and thinking. “The biggest complaint is that the candidates don’t feel that it was like a real human being talking to them.” AI may not be about to replace interviewers, “But it does give them a superpower,” says Tajbakhsh. “It’s an augmentation. But you still have to be able to execute.”

In bygone days, if I was ever going for an interview, my mother always used to say, “Just be yourself, dear.” I was always pretty sceptical about that advice. I finally worked out that the trick (I offer this secret for free) is to keep your hands on your knees: it’s good for posture and it stops you flapping. Does Metaview have any tips?

Siadhal Magos’s very useful advice is: “As a candidate you don’t want to come across like you’re running from something – you’re running towards something.”

“Maybe your mother was right,” says Shariah Tajbakhsh. “It all comes down to working out what makes you tick. Successful candidates are the ones who have found their zone of genius. It’s pointless hacking your way into something you’re not good at. If you get in it’s not a hack, it’s the truth.”

@andymartinink