By Ben Pfeifer, Limelight CEO
The hiring process for many jobs is fundamentally flawed.
There’s too much emphasis on credentials and pedigree, and too little on evaluating the actual skills that determine job performance. Instead of using a calibrated, data-driven analysis of candidates’ hard and soft skills, the hiring process often relies on a few unstructured conversations followed by a subjective hiring decision. We are solving this problem at Limelight by creating a skills-based hiring platform that allows candidates and employers to find the best jobs and the best talent possible using a structured, data-driven assessment process. We are currently focusing on careers that are built around specific learned skills, including manufacturing, construction, and the trades. Our approach uses data to produce more successful hiring outcomes while eliminating much of the subjectivity, bias and discrimination embedded in pedigree-based hiring. It is also the only approach that is sufficiently fast, accurate, and cost-effective to meet the challenge of getting our economy moving quickly again as COVID-19 recedes.
Here’s how we’re approaching this challenge.
In 1998, Tim Westergren was thinking about how to connect musicians and audiences when he came up with the idea of creating a taxonomy for music. If he could define the ‘genetic code’ of a piece of music, he could help an artist reach new fans accurately and efficiently by analyzing music they already liked. This effort, which Tim dubbed the Music Genome Project and has become the most comprehensive analysis of music ever undertaken, now categorizes songs using over 400 attributes including genre, vocals, and emotions. The project created a framework which brought order to what seemed to be the impossibly subjective world of music, and allowed for precise matching of music to listener, at scale. The Music Genome Project powers Pandora, one of the world’s largest music services.
Like music, jobs may seem to have an infinite number of unique attributes, but that belies the large degree of commonality among skills required for a particular position. That’s why we think of our detailed skills taxonomy as Pandora for jobs.
“…we think of our detailed skills taxonomy as Pandora for jobs.”
With a job’s skill taxonomy in hand, how do we find the best person to fill the role? It turns out that understanding the applicant side of the job market also has an analogue. About a decade ago, Dan Cederholm and Rich Thornett wanted to create a better way for visual artists like designers to showcase their abilities. They created Dribbble, which allows artists to create image-heavy profiles that are much better at capturing style and skill than a resume ever could. It’s the same thing for many other jobs: A list of machinist positions on a resume provides only circumstantial evidence of the applicant’s skills. Now imagine an opportunity for that machinist to prove their knowledge during the hiring process.
“… we think of our Limelight assessment as Dribbble for skills.”
Interestingly, applicant testing used to be common among U.S. employers, but that practice was challenged by several landmark Civil Rights laws and court rulings that found that many of the tests, as implemented at the time, amounted to discrimination. That’s a key reason why so many employers retreated from assessment testing to the superficially non-discriminatory approach of credential-based hiring, with particular emphasis on the college degree. It’s a key reason why the college degree has gained outsize importance in the current job market. In fact, a study published by Harvard Business School in 2017 found that for a middle skills job title like production worker supervisor, 67% of job postings required a bachelor’s degree, although only 16% of those already in a comparable position actually had one. And even more surprisingly, those without the degree actually performed better in those jobs.
In recent decades, colleges have taken full pricing advantage of their position as the increasingly unquestioned gatekeeper to a large percentage of good jobs. Since 1980, tuition and fees at four-year colleges and universities have risen 17 times faster than average family incomes, according to the National Center for Education Statistics and the St. Louis Fed. The resulting student debt crisis has been well documented. College is still a great investment for many — it’s just that excessive credential filtering has turned it into an expensive hurdle that unfairly excludes talented people from their optimal careers.
“College is still a great investment for many — it’s just that excessive credential filtering has turned it into an expensive hurdle that unfairly excludes talented people from their optimal careers.”
Other countries have well-developed vocational education systems that prepare workers for technical jobs. In Germany, for example, legally-mandated cooperation between government and industry intertwines education and on-the-job training, with nationally standardized testing and certifications. This system has been a crucial underpinning to Germany’s vaunted Mittelstand industry, which includes thousands of small and medium businesses producing items including machinery, auto parts, and electrical equipment that are in high global demand. We can take the best of this approach and adapt it to the more flexible, free-market culture of the United States.
Consider the current process of filling a job. Whether manually or using an automated system, an employer scans incoming resumes for keywords and credentials that make a candidate seem worthy of a closer look. Circumstantial evidence carries outsized weight, and often only a small number of candidates make the cut. Even if hiring managers give skill tests to the candidates that made it through the initial screening, the game has already been misplayed, if not lost. Good candidates with ineffective resumes or missing credentials may have already been discarded, and the hiring manager will also likely spend a lot of time with unsuitable candidates.
With Pandora-for-Jobs and Dribbble-for-skills, Limelight is creating a matching process that is accurate, substantially automated, and immediately relevant for the specific job at hand.
“With Pandora-for-Jobs and Dribbble-for-skills, Limelight is creating a matching process that is accurate, substantially automated, and immediately relevant for the specific job at hand.”
Given that use of skills assessments was reduced out of concerns of bias, it’s interesting that companies that have introduced modern assessment tests have found that they increase diversity. It’s an indication that despite all conscious efforts to eliminate bias in the hiring process, nothing succeeds like a neutral, skills-based test. In 2015, Ernst & Young removed degree requirements from their job descriptions and added assessments. They saw a 75% increase in applications and have become one of the UK’s top ten employers for diversity and social mobility.
Jobs built around specific learned skills are a huge and growing part of the American job market. The Georgetown University Center on Education and the Workforce published a study in 2017 that identified 30 million “good jobs that pay without a BA,” up over 10% from 1991, that paid an average of $55k/year. Many of the identified job categories were growing quickly before COVID-19, like healthcare and construction, and can be expected to resume their strong growth once the economy shakes off the virus.
The silver lining of COVID-19 is that the virus has made reform unavoidable. We are building a fast, accurate, and cost-effective way to re-hire millions of workers, efficiently guiding them to new opportunities. There will be an unusually large amount of hiring in the coming years, and we believe that fundamental improvement of the job market is required to help the country rebuild, while improving diversity and remaining globally competitive. Join us.
About Ben: Ben Pfeifer is the CEO of Limelight. Limelight is backed by 25madison, a leading venture and investment firm based in New York City.