Simple questions like 'do you like horror films?' can predict whether a startup will implode

London-based "people analytics" startup Saberr claims its 70-question quiz can determine how well a team will work together - and even if a startup will fail or not.

Founder Alistair Shepherd told Business Insider: "[Early stage venture capital fund] Seedcamp are one of our own investors.

"They had a shortlist of 21 companies they were considering investing in and we said to Seedcamp a) we know who you're going to invest in and b) we know how they're going to fare. One of the partners quite rightly said 'Nonsense, but I'm intrigued so run your algorithm.'"

Saberr gives each member of a team a 70-question quiz to determine "your values and your tolerances," according to Shepherd.

"We ask fairly off-beat questions like 'do you like horror movies?'," he says. "In this particular question, you're looking for agreement. Your opinion on horror movies, statistically speaking, is a very valid question to ask."

Saberr then crunches the data through its algorithm to give an idea of how well the people in that team will work together.

"Right now there are lots of personality tests you can do out there that can tell you about your personality," Shepherd says. "But that's not particularly interesting as it doesn't tell you how you interact with other people."

The tests on the Seedcamp investments revealed one company that Saberr thought would fail. "We knew nothing about the skills, the demographic or what they were working on," Shepherd says. "All we could see was they had a fairly mediocre dynamic between three of the four people and there was one character we thought was going to cause issues.

"One week after the paperwork was all signed [on Seedcamp's investment] that company folded for the reason we had suggested."

A similar scenario has played out on a number of other occasions, such as the time Saberr correctly predicted all but two of the top 10 winners in app building competition Microsoft Imagine Cup. (They got four and six the wrong way round.)

"My background is aerospace engineering," Shepherd says. "I come from a very clinical background. Unless you can prove something to be right, then it is forever wrong in an engineers mind. We focused very heavily from the beginning on how do we prove what we're doing works.

Investors use the three-year-old startup for due diligence on companies they're looking to put money into, but Shepherd says the main use cases are for hiring and internal team building.

"Our focus is on people with a 100 people plus," he says. "It works at the startup level but often what we find is founders don't want to hear the bad news."

That said, Saberr use their own formula when they're hiring. "Nik Brbrora our CTO, we predicted his startup would fail and we also predicted he would do very well in our company," Shepherd says. "When that prediction came true, we asked him to join our company."

Saberr raised $1.2 million last June (it used its test on prospective investors) and is currently raising another $1.5 million. Customers include Big Four accountant Deloitte.

Saberr is also working on adapting its predictive technology to social media. Shepherd says: "In future you might not have to answer questions like this but you can just submit your Twitter profile or your LinkedIn profile. Your Twitter profile has a rich depth of data on you that tells me a lot about you."