Algorithm Alchemy: Turning Talent Search Into Gold

Algorithm Alchemy: Turning Talent Search Into Gold

How much do companies spend on people? Trillions of dollars… Winning or losing your market is based on whether or not you have not just good, but amazing people.

-Dan Shapero, VP of Talent Solutions, LinkedIn
As of today, no one’s found a scalable* way to provide general recruiting options. So far a cadre of startups have tackled recruitment, but focus primarily on the recruitment of engineers, programmers, and coders.
VLAB’s June panelists framed the problem of the job market as signal versus noise. Hundreds or thousands of applicants overwhelm the most seasoned recruiters, and job alert emails become spam in every searcher’s inbox.

What if an algorithm could cut through the junk, and deliver the perfect candidate to a business? Algorithms outperform human hiring decisions by 25% according to Craig Driscoll of Highland Capital. And what if the job were so compelling that the perfect candidate couldn’t help but apply?  If you can introduce the perfect candidates to the perfect companies at scale, you’ve got a revolutionary product.
1) Let’s dissect the problem
Life’s hard for both recruiters and candidates.
What’s making recruiters’ lives so difficult?
Discoverability. It’s too noisy. “It’s too hard today to get somebody to work for you,” said Alex Dévé of Whitetruffle. Marc Hoad concurred, “You shouldn’t be able to apply to every single job on the planet.”
Sheeroy Desai, CEO of Gild, points to “this whole idea of quality… How do you find quality at scale? … The criteria of shortcuts that companies are taking today do not have a very high correlation to
[successful hires].”
What’s more, the perfect candidate might not know he wants a job. Shapero broke the employment pool into three categories:
20% of people actively look for work. These are the people who send in their resumes. “They’re either out of the job or they hate what they’re doing. 60% would love to have a conversation if you show them the right thing.” A friend of mine fits in here. He’d like a change, but he’s too busy to look. 20% “Couldn’t be bothered”; you’ll never get them to budge.
To solve the recruiter’s problem, expand your data pool to include the 60% in the middle while filtering through the noise to make sure you find exactly the right candidate.
What’s making candidates’ lives so difficult?
The 20% of candidates who actively search must compete with the noise of other applicants. Oftentimes, resumes land in the wrong bin (the recycle bin, if you were wondering) for completely irrational reasons: decision fatigue, intuitive bias, etc. A colleague and I once gave a candidate an interview because we caught a joke on his resume. Do jokes on resumes correlate to successful hires? You might as well say he got lucky.
All candidates suffer from the noise of irrelevant job listings. I’m still unsubscribing from job boards I joined a year ago; they come so often I don’t look at the postings. And what’s to say a particular job board actually lists the perfect job for a candidate? How do you expand your data set to include all relevant jobs to send candidates only the perfect fits, cutting through the noise?
Lastly, candidates can be oblivious. “No one knows what they want to do… The lack of data in people being able to choose their careers is a data problem,” said Shapero.
2) Current solutions from the panel
Gild’s Sheeroy Desai: “Our hypothesis going in is that there’s an incredible amount of talent out their that’s getting ignored… [So] we started leveraging public data… There’s a richness of data in there that can give true signals of quality.” Gild mines open source coding communities to look for quality engineers. In doing so, they bypass traditional- and ineffective- systems that sift candidates by pedigree, seniority (or youthfulness), etc.
Alex Dévé of Whitetruffle “was trying to figure out what to do with [his] life besides watching soccer games…” and so created a dating site for jobs. He has a matching algorithm that sends candidates to recruiters, and a delivery algorithm that sends specific job openings to candidates. “There is no spamming. It’s not possible.”
Marc Hoag’s Crowd Vitae helps candidates “perfect” their resumes, helping them rise about the noise. Venturocket actually connects job seekers directly with hiring managers based on skills and expertise fits. “We wanted to eliminate the resume spam problem… Essentially a Quora for resumes.
And LinkedIn. Speaking for the world’s dominant online professional network, Dan Shapero said, “Our answer is to open up the world to the 80% of talent that aren’t looking for work.” Recruiters, a key subscriber base for LinkedIn, can find all those people out there that don’t even know they want a new job.
3) Some killer-app dreams
1) Imagine if you could click on your dream job, and then view the average career path of 50,000 individuals who eventually got it. LinkedIn is slowly amassing the data; can you analyze it?
2) Virtual career mentorship. Dévé elaborates: “Imagine if a [college student] could sit down with us and say, ‘should I take this elective? Should I take an internship at Facebook or join the peace corps?’” No more college students advising college students; everyone gets advice tailored to their context.
3) Harvard business review talks about potential versus experience. But how do you measure potential? And how do you quantify soft skills and compare to a universal baseline?
Craig Driscoll pointed out that funded companies in the recruiting industry have frequently been acquired or gone public, “which is encouraging…”
(Written by Eric McClellanVLAB Marketing Committee Blogger. Eric coaches start-ups in UX design and external communications.)
Other Related Articles by the VLAB Community about the Event:
Creative Minds In and Out of Silicon Valley | Dr. Katherine Jones, 2014-JUL-09
The Young Guns of Talent Acquisition Tech | Martin Shaw, 2014-JUL-09
By | 2017-05-18T17:47:12+00:00 July 16th, 2014|VLAB|Comments Off on Algorithm Alchemy: Turning Talent Search Into Gold