A recruiting approach that uses algorithms to weed out resumes is inherently flawed, because it necessarily relies on such a narrow set of data. Instead, the power of algorithms and machine learning needs to be tapped as part of a holistic approach that enables recruiters to efficiently consume large quantities of data in order to make the best hiring decision possible.
That was my key takeaway from a recent interview with Aaron Elder, co-founder and CEO of Crelate Talent, a cloud-based human capital management platform provider in Kirkland, Washington. Elder, a former lead CRM software design engineer at Microsoft, sees his startup as focusing not just on the technology, but on the human dimension of hiring. “Crelate is the mixture of two words: create and relationships,” he said. “So we’re creating relationships, and we’re nurturing and growing talent.”
I asked Elder to define the problem that Crelate Talent wants to solve, and he explained it this way:
While I was building consulting businesses, we were constantly hiring people and placing them on projects. Where my heart sort of landed was on aligning people with opportunity, because when that happened, I always saw people’s careers skyrocket, I saw good things get produced, I saw happy customers, happy people, everyone benefitted and cool work got done. And so when it came time to start a new venture, I wanted to build upon all this, and actually build a CRM that’s focused on selling people, as opposed to selling widgets, which is where Salesforce and [Microsoft] Dynamics had their roots. And so we set out to build an applicant tracking system and a recruiting CRM from day one to do that. There’s no shortage of players in that space, so how are we different? The way that we want to position ourselves is that we actually consider the role of the recruiter to be a noble profession. It’s as old as human society — it’s matchmaking, and we don’t see this going away. So we want to build tools that enable these people to make better decisions and spend their time more effectively, not replace them. I think there’s a trend emerging where machine learning and analytics are kind of a new snake oil. The idea is that you don’t really need recruiters — the algorithms will tell you who to hire, and so on. Instead, we’re building tools to help good recruiters be great recruiters — that’s our goal.
Elder went on to explain that that faith in algorithms might me misplaced:
I was reading an article the other day about how using algorithms will build a better workforce, and their recommendation was, let the algorithms weed everybody out, and then have your recruiters make the final decision. The issue is that 72 percent of resumes never make it to a human, and I think that’s inherently flawed, because ultimately, there are a few problems. One, algorithms, unfortunately, can reinforce the bias of the programmer. Two, I think there are plenty of people who just stink at writing resumes, and so you’re bound to miss lots of good candidates. Three — and I think this is the biggest point — it’s not an easy problem to solve, and the first step is knowing the problem. If you think about the entire wealth of your career, your experience, everyone you’ve known, every good thing you’ve ever done professionally, and then you think about what percentage of that is actually captured by a resume, it might capture 1 percent, or if we’re generous, 5 percent. The best matching algorithm in the world, based on such a narrow view of the problem, can’t be that great. So I think this first generation of tools that are attacking it from that angle is basically going to be drawing conclusions that are not so accurate.
Elder said Crelate’s alternative is to use algorithms and machine learning differently:
You want it to make suggestions. You don’t want it to give you answers, and you don’t want it to filter stuff out. You want it to make it easier for you to consume big pieces of data quickly and efficiently. So instead of never seeing the stuff, it lets you see it, but it also lets you efficiently weed through it based on what you want to do. It’s all about allowing recruiters to make more effective use of their time, and to uncover those diamonds in the rough. Ultimately, software is designed to help save time. Far too often, software actually ends up becoming the work itself. We are trying to build software that doesn’t make work for you. I don’t like the idea of a shiny ball or one magic answer to all your problems. It might have made a good headline, but ultimately I think human relationships are complex. The hiring process is important, and I think a company needs to have a more holistic view of things to be successful.
Elder wrapped up the conversation by summarizing where he sees Crelate Talent heading, and why:
When I left my last company, I left a large, people-centric business. The primary asset of this company was people, but it couldn’t have been any more cold or heartless. If your company’s primary asset is people, you would think that the software and tools that you use are all about enabling and empowering those people. Let’s build software that helps employees make better decisions, and be more successful. That helps everybody.
Today, we’re primarily a recruiting CRM and applicant tracking system. Two years, five years out, I want to be using machine learning and software to solve this problem of matching by enabling professionals, but doing it on a much larger set of data than the 5 percent that a resume or job description could ever provide. This is where it depends on what LinkedIn is going to do with their social graph. Certain players have access to the data you would need to actually get that larger picture. As an individual start-up, I don’t have access to that right now, but someone like Microsoft does. Long term, I’m hoping that the powers that be enable access to these graphs to enable all kinds of new solutions to be developed outside of their walls.
A contributing writer on IT management and career topics with IT Business Edge since 2009, Don Tennant began his technology journalism career in 1990 in Hong Kong, where he served as editor of the Hong Kong edition of Computerworld. After returning to the U.S. in 2000, he became Editor in Chief of the U.S. edition of Computerworld, and later assumed the editorial directorship of Computerworld and InfoWorld. Don was presented with the 2007 Timothy White Award for Editorial Integrity by American Business Media, and he is a recipient of the Jesse H. Neal National Business Journalism Award for editorial excellence in news coverage. Follow him on Twitter @dontennant.