Companies as varied as Deutsche Bank, Airbnb, and Reebok are turning to Seattle startup Koru to help them recruit graduates.
Koru, founded in 2013, uses machine learning and predictive analytics to predict whether candidates will be the right "fit" for a company. Applicants are made to take a 20-minute online test that measures seven personality traits: grit, rigour, impact, teamwork, curiosity, ownership, and polish. It dubs these the "Koru 7."
Your results are then matched up to the company's to see how your specific qualities would likely contribute to the organisation.
CEO and cofounder Kristen Hamilton told Business Insider: "We started the business really fixing a problem as opposed to developing a product. It was really hard for people to pick who to hire when people had the same experience."
"A typical company when they get a huge number of applicants, they slice down that first level of applicants based on two things -- where they went to college and how their grades are. That cuts out a lot of potential high performers. Those 2 things have really been proven to not be terribly predictive of future performance. It also cuts out a lot of diversity, so a lot of people from different backgrounds don't get an opportunity to do well."
Hamilton and cofounder Josh Jarrett jumped on a video call from Seattle, where Koru is based, to talk Business Insider through the test and how it works. They also provided BI with some mocked-up, dummy images of what the tests might look like so you can get an idea of how it works.
Before a job applicant sits down to take the test Koru has to do some work with the company. Its test is not a one-size-fits-all solution and is tailored to each company that uses it.
Hamilton says: "We recognise that each company is unique. What it takes to perform well at Deutsche Bank is different to what it takes to perform well at different investment banks we work with, even if they're in the same industry and in the same job."
Aside from Deutsche Bank, Koru works with clients like Citibank, Reebok, Airbnb, McKinsey, and others.
Hamilton says: "We first do this thing called 'fingerprinting,' which is where we go into a company and take a read on their current employees to figure out what drives performance. For companies that's useful because sometimes what they think drives performance is wrong and sometimes it's additive."
Koru makes current employees take its test, looking at poor, medium, and high performers so that it can tell not just what it takes to get into a company but also what it takes to do well there.
The test is then tailored to what the company in question is looking for in prospective employees -- it's not always just about finding top performers.
Jarrett says: "In the fingerprint we define what the important metrics are for the organisation. It could be we're looking for performance in the first 2 years, it could be that we're looking for retention, it could be that we're looking for future leadership.
"The process of fingerprinting uses machine learning algorithms to build those predictive models. Knowing that curiosity is one of the things that drives high performers doesn't mean that you should hire everyone that's curious."
Hamilton adds: "In Silicon Valley they're getting very serious about diversity. There's a lot of data showing diverse teams are more effective. The customer base is so diverse and to really serve that customer base you have to reflect that diversity. That's very serious."
Without further ado, let's take a look at the test...
Here's what you see when you start the test
Once the fingerprint is set up, the company can apply the Koru test to all its vacancies.
When potential interviewees submit their resumes they'll be sent an email with a link to the test.
"It's pre-interview and it's really a chance to get to know you better," Jarrett says. "We know you're more than just a resume. We'd love to interview everyone face to face but since we can't we'd like to ask you some more questions about your past experience, your background, etc."
The test isn't timed and can be left and returned to later, but if you were to do it in one sitting it would take roughly 20 minutes.
Stage one: Past experiences
The test is split into three sections, the first of which is past experiences. This is all about looking for examples of when applicants have used certain skills in the past.
Jarrett says: "If you've used the skill in the past, you'll be better at using it in the future. Looking at past school activities, extra-curricular activities, committee activities, jobs -- not just that you did something, you were on the Volley Ball team. That's not helpful.
"It's what you did underneath that -- you managed the trip to France to play a team there, it was a big project management effort and you had to manage a budget. You were a shift manager at McDonald's -- you had to get up early, you had to motivate people. Those are the skills that will be transferable."
Step two: Work style
In the work style section, where candidates are asked to choose which of two statements sounds more like them.
Jarrett says: "Would you rather solve a hard problem in Excel or would you rather solve it with a team in a team room and a white board. You can imagine that if you prefer the first you've got more analytic rigour, but if it's the later you're more about teamwork, collaboration.
"We're looking for those trade-offs, we believe that not everybody is great at everything. Where are you strengths and do those strengths line up with the companies strengths."
Applicants will have trouble trying to "game" the Koru test as even if you can guess what qualities the choices relate to, they won't know what qualities the company is after in its fingerprint.
Step three: Work scenarios
After the work styles section comes work scenarios, a multiple choice section meant to test how people would judge situations.
Jarret say: "We give you a short scenario: you worked really hard, you pulled an all-nighter, you handed in your report to your boss but she didn't even look at it, just put it in her draw and didn't even say thank you. What do you do?
"Do you put your head down and keep working hard, do you go to her boss and complain, do you schedule time with your boss to talk about it, do you go out with your mates for a beer and complain about what an awful boss you have?
"You can imagine the different answers signal someone with high emotional intelligence, someone with high self-efficacy."
By the end of the test, Koru has built up a profile of you based on its own unique profiling metrics.
Hamilton says: "We come with a point of view, we have the Koru 7, that's our IP, we've developed that. That's our point of view of what millennial talent and companies need."
This point of view has been backed up by extensive research into the effectiveness of the "Koru 7" in predicting future performance.
Koru's predictive model should also theoretically be getting better and better the more people take the test. Hamilton and Jarrett wouldn't disclose specific numbers but the Financial Times reports that 30,000 people have taken its test.
Every candidate who takes the test then gets a personalised profile showing them their biggest strength. Koru also offers advice on how to improve some of the areas where they came up weaker.
The benefit is that candidates who don't even make it through to the interview stage still feel like they've got something out of the application process.
Jarrett says: "The applicants receive feedback, they get a high-level report on their top skill and how they can use that. The employers we work with really believe strongly in applicant's experience. Those candidates show up on a dashboard for recruiters and that's arrangeable and exportable."
...and here's what the company sees
For recruiters, they can see all the applicants for a particular vacancy on one portal. They can be sorted by lots of different metrics but, most importantly, can be filtered by match to Koru's company "fingerprint."
Jarrett adds: "There's multiple factors that go into it and human brains have a hard time of holding multiple weighted factors for every candidate they see. That's what our machine learning algorithm can do."
The Koru portal also gives tips on how to ask questions exploring qualities that are shown to work well in your organisation, like "Grit" for example.
There's then an "after care" element to Koru's platform. For companies, Koru tracks the performance of those people you hire based on its recommendations. This feedback helps to improve and refine its algorithm.
Jarrett says: "Whether or not you interview someone is the first thing that makes the algorithm smarter and then the second metric is performance -- from 90 days in the case of sales to 12-18 months."
Different companies use Koru for different things so what it measures differs. Jarrett says: "Retention is certainly something in finance -- where is our future workforce coming from? How are investment banks not just a training ground for private equity firms?"