Everyone -- and no one -- knows the solution to the job recruitment conundrum
The other day when I logged into Facebook, it prompted me for more information about myself. I've always been a lazy Facebook-er, never the type to fastidiously detail my likes and dislikes, brand preferences, and favorite books for the masses. My job history was spotty. I always figured anyone who wanted to learn about me would get more from reading my wall than my bio.
But apparently Facebook wasn't satisfied with that. In the top left-hand corner under my profile picture, it prominently displayed a q&a. Only seven questions. Easy enough.
It was hard to ignore. Who doesn't want to answer a quiz about themselves? Narcissism, man. It's motivating.
When did I work at City 7 Television? What job did I hold at the age of 16? Who are my favorite musicians? If I couldn't come up with any answers I could choose from a list of prompts below.
Of course, Facebook totally tricked me, and after I finished the initial seven questions, 30 more popped up. At that point, I was in the flow and compulsively wanted to wrap the whole thing up, so I put up with the quiz and filled out my profile for the first time since I created a Facebook account seven years ago.
After I finished I had two thoughts. One: Well done, Facebook. Slow clap. Your behavioral trick elicited more information out of me for your advertisers. Two: This bodes well for Identified, the first startup I ever met when I arrived in Silicon Valley four months ago.
Identified, which launches today, is trying, like so many others, to fix job recruiting through a data driven approach. I got an introduction to their technology, called SYMAN, back in June during an interview with Identified's angel investor Bill Draper and co-founder Brendan Wallace.
Identified scrapes publicly available information from people's Twitter, Facebook, and LinkedIn profiles to create a giant database of job candidates. It anonymizes the user information, so that when a recruiter searches the database and finds a potential candidate in the right location, they need to go through Identified to contact that person.
The system that powers Identified's database, SYMAN, uses semantic algorithms to makes sense of the "messy" public data. That way, if a recruiter is searching for a hedge fund analyst in D.C., they don't accidentally pull up CIA analysts.
So why wouldn't recruiters just use LinkedIn for such searches? "A nurse, a policeman, a store manager, are unlikely to build a LinkedIn profile," says Brendan Wallace. "So we enable recruiters to find on Facebook where people put that information." Facebook isn't exactly easy to search however -- a search for nurse may very well pull up a woman named Nurse in addition to those with the career of nurse. So SYMAN cleans up that data, organizes it, and makes it easy for recruiters to sort through.
Identified's value add is that it can surface people with unique backgrounds and experience that recruiters might otherwise struggle to find. For example, it's unlikely that a recruiter just looking for a 'nurse' would need to use Identified. They would probably receive an abundance of qualified applications just by posting the job online. But for highly specialized roles, perhaps in less-populated areas, Identified could be really helpful.
"For example, Yale-New Haven Hospital had a role outstanding for 15 or 16 months that they couldn't fill. They ran a search through Identified, and found a therapeutic recreational specialist in New Haven Connecticut," Wallace says. "It was hard to find, but we were able to build that information and connect them."
Of course there's a couple problems with scaling this strategy, first and foremost that I don't believe huge amounts of the population extensively list their career background on Facebook. In fact when I'm shamelessly Facebook stalking people from my past I frequently try to figure out what they're up to. More often than not, their job information isn't listed.
Wallace disagrees with me. "It's really not true. There's a tremendous amount of professional information on Facebook," he says. "It's not the same depth of information as, say Linked-In, but it will have the location, job titles, years you worked."
Facebook's latest prompts will greatly benefit Identified if they trigger other people to fill out their profile. More completed profiles take Facebook one step closer to fulfilling its social graph mission and becoming a one stop shop for information on individuals around the world. It's a social graph that companies like Identified can capitalize on.
But the second problem Identified's strategy poses is scaling its algorithm.
To identify hard-to-find candidates, Identified semantically broke down healthcare jobs. For example, it distinguished between the role of an ER nurse versus an end-of-life-care nurse. People manually taught the algorithm how to identify qualified candidates. As you can image, the process took time -- 3-4 months.
At the end of it, Identified now kicks butt at surfacing qualified job candidates in healthcare through messy social media profiles.
But what about all the other industries? With Identified's launch today, you can search any industry for job candidates, but you may not surface those hard-to-find people. The company hasn't yet done the "deep semantic dive" for other job verticals.
That will change soon. Identified is tackling retail and consumer products now, and when it's done it will move onto other fields.
I told Wallace it sounded like this approach would take a lot of time to scale across different industries. He begged to differ. "Job titles have a lot of similarities across industries in terms of compounds and structure, so we've been able to now move at a faster pace," Wallace says. "It took three to four months for healthcare, but retail and consumer products we're doing in the course of weeks."
I hope he's correct, given I can't help but have a soft spot in my heart for the first startup I ever learned about at Pando.
That said, the data-driven job recruitment scene is a competitive one to break into. GroupTalent, WhiteTruffle, Apploi, Bright, TalentBin, Jobandtalent, it seems like every week there's a new company launching claiming to fix the problem.
Some, like Bright or WhiteTruffle, collect resumes so they can run their own algorithms for job matchmaking, but are then limited to the resumes that are submitted to them. Others like GroupTalent or TalentBin, focus on specific verticals like matching designers or developers to jobs.
Identified might work well in its own niche, of helping employers of industries underrepresented on Linked-In with hard-to-find candidates. It also fits well in the increasing trend of employers using social profiles to vet job candidates.
But I'm not sure it will be able to dominate the data-driven job recruitment scene. Its approach relies on the unpredictable external factor of people filling out their social media profiles with extensive career information. And even if people did that, I could only see a recruiter using Identified to fill hard-to-find roles.
The job recruitment startup scene is increasingly fragmented. I can't help but wonder if in the fight to the finish line, they're all going to die?