This Mark Cuban-Backed Startup Invented a Game That Could Completely Change the Way You Hire
Founded by a former White House staffer, San Francisco-based Scoutible is using artificial intelligence to change the hiring process.
PHOTO CREDIT: Getty Images
Hiring someone who turns out to be a bad fit can be costly: Unhappy employees cost the U.S. economy between $450 and $550 billion in lost productivity each year, according to research firm Gallup. And replacing a full-time worker can cost up to twice the employee's salary.
While working on a project at Harvard Law School, Angela Antony found herself immersed in statistics like those.
"If you look across the economy, about 46 percent of hires leave within 18 months. That's despite all the time, resources, and billions of dollars spent trying to effectively hire," Antony says. "My research was really trying to understand what that missing data set was that was preventing us from being able to accurately predict people's long term performance in a role."
That research helped land her a job at the White House with the National Economic Council in 2015. It was there, at an entrepreneurship policy event, that she met Mark Cuban. Antony told him about her research and the book she was writing on the topic.
"Mark said, 'Don't finish that book,' " Antony says. " 'You should build the solution, and I want to fund it.' "
The result is Scoutible, a San Francisco-based startup that wants to take the mystery out of the hiring process. Scoutible's product isn't an exam or a questionnaire--it's a video game. During 20 minutes of gameplay, the system collects millions of data points used to measure a candidate's various attributes, like problem solving abilities and risk aversion. It then produces a numerical assessment of how likely he or she is to excel at the role in question. Antony says the insights are far more effective than traditional screening methods, such as resumes or interviews when it comes to assessing a good fit. Cuban has invested more than $1 million into the company, which is announcing its $5 million seed round Wednesday.
Gaming your way to a job profile
When you play Scoutible's game, you take control of a character who is sent on a series of missions, such as surviving on a deserted island or thwarting an assassination attempt against your fantasy world's king. Your various decisions, including the way you navigate the map or interact with computer-controlled characters, give Scoutible's artificial intelligence a fingerprint of your underlying attributes. For instance, someone who falls through a trap door might show an aptitude for dealing with minor crises by quickly recovering and continuing with their mission.
Antony, who studied psychology as an undergrad at Harvard, was inspired to go the gaming route when she recalled the various tests of cognitive traits she'd encountered in college. "They very much felt like games," she says. "I thought if we could build a tool that felt like a game, and incorporated tests for the appropriate personality and cognitive attributes, we could address many of the problems I'd been studying."
The startup's 10-person team, largely made up of psychologists and engineers, built the game and its underlying A.I. from scratch. It's designed to measure factors like grit, interpersonal skills, leadership ability, and creativity--attributes that a resume or job interview might not effectively communicate.
The game's purpose is to give employers an advantage when it comes to identifying candidates who are the best for for their particular company and role. In most cases, the game will act as the first step in the process for a job seeker, helping employers focus on people whose results are in line with what they're seeking. Antony says that there's no way to be "good" or "bad" at playing the game; instead, the players' decisions determine whether their personality and aptitudes match the job description.
For certain roles, like salespeople and customer service representatives, Scoutible has established baseline metrics that can then be lightly tweaked from employer to employer. For positions with less established blueprints for success, Scoutible works with the company to collect data from workers who have excelled in the same or similar roles in the past. It then creates an ideal profile from which to measure candidates. The gameplay and the system's algorithms do the rest.
Outsmarting your own bias
"Self-reported personality surveys are usually not accurate when used in a competitive context like job evaluations," Antony says. On those, candidates can easily shape their answers to fit what they think an employer wants to hear. Other times, the inaccurate self-reporting is less deliberate: A candidate might truly believe that, if hired, he'll be able to change his behavior to match what the company desires.
"Our game solves those problems," Antony says. "It's a type of stealth assessment. At no point in the game is it ever obvious what the desired answer is, or how to represent yourself in a way that elicits some outcome that you think employer wants. The pacing and the situations are such that you have to just sink into the game and act based on your most authentic and instinctual responses. You're just reacting to things as they occur."
Like with any form of A.I., there remains the risk that Scoutible's system could perpetuate the biases of its creators. But Antony claims it will have the opposite effect. "We basically allow people to cut through the noise based purely on the candidate's ability to do the job well," she says, "without ever really exposing the company in the evaluation process to those biasing factors," like gender or race.
So far, Scoutible has about 20 clients, ranging from startups to large corporations, including tech companies, retailers and hedge fund managers. "It seems this is a pain point people recognize," Antony says, "and want a different solution for."
The results, she says, have been promising. For one 101-person sales department, Scoutible's game predicted job performance with 4.5 times the accuracy of traditional job interviews. On another 150-person team, it predicted representatives' customer service ratings within 1 percent accuracy.
Eventually, Antony says, the startup wants to roll out more consumer-facing features, though she declined to go into specifics. For now, the company is focusing on growing its team so it can handle more clients, and building the strongest possible version of it's A.I.
"We want to place people in roles," Antony says, "where they'll not only outperform, but where they'll also be the best culture fit and have the most rewarding long term experience."