Meet a Start-Up With a Big Data Approach to Hiring
A recent survey from Silicon Valley Bank asked start-up founders and technology executives to name their biggest challenges. About 90 percent listed "finding talent" among their chief concerns.
It makes sense. In Silicon Valley, the life and death of your start-up can literally depend on your earliest hires. Unfortunately, the process of finding them--using recruiters, job boards, LinkedIn, or even word of mouth--isn't always easy.
Alex Deve, the co-founder and CEO of Whitetruffle, an 8-person start-up in San Francisco, thinks he can change that. How? The answer, he says, is in the cloud.
"Our vision is to become almost like a Pandora for hiring," says Deve. "We can serve both candidates and companies what they really like."
Whitetruffle describes itself as a "tech talent-sourcing platform." Candidates enter information about themselves, and companies enter information about the type of candidate they're looking for. Using a proprietary algorithm built by Deve, Whitetruffle's technology scans about 50 categories of "signals" in a candidate's profile. Then, it scans the job requirements of company postings to find a match.
That may seem simple, but the elegant part of Whitetruffle is in its subtleties of data design, something Deve calls "intent." The mechanics of intent, however, may be a more difficult piece to the puzzle. It will require fine-tuned algorithms--and plenty more data before it can be truly effective.
Learning More From the Data
"First of all, the algorithm is a work in progress," he says. "We've made great progress. We are currently looking at more than 50 different 'hard-core' signals, which means there are hundreds of combinations possible. There's the usual suspects--location, skill, employment history, education. But that's just one piece of the puzzle--and that piece is actually not difficult to build."
"The really hard part," he says, "is when you try to combine relevancy with intent--that's a very important point in what Whitetruffle is trying to do."
Intent is a tricky concept, but here's how it works. A few hours after a company joins Whitetruffle, they are given a list of candidates for a particular job. They simply say 'no' or 'yes' to whether or not the company would hypothetically be interested in the candidate. All the while, Whitetruffle is recording those responses and building a deeper understanding of the type of candidate that company is looking for.
"That's actually very interesting data to have," Deve says. "For example, if we have a candidate in the system and we show that candidate to 10 companies, and the 10 companies want to talk with that person, that tells us a lot about that person. Then, we look at what those candidates say about those companies, and that shows us a lot about that candidate and the companies. We manage all that data to actually get better at matching."
In other words, the more data Whitetruffle gets, the smarter it becomes.
Since its beta launch two years ago, the company claims 30,000 job-seekers have signed up for the platform, and 2,000 companies have used it to look for potential candidates. According to Deve, managers have cut down on sourcing time by 30-40 percent, and about 50 percent of all hires made using Whitetruffle came from candidates who already held full-time jobs.
Though the company won't release precise statistics, Deve says Whitetruffle has successfully contributed to "hundreds" of hires at several hundred fast-growing tech companies, including Asana, Pebble, Eventbrite, LearnVest, and Dropcam.
Up until now, Whitetruffle's revenue model has been like any typical recruiter--it charges a fee based on 20 percent of the candidate's first-year salary. But in an effort to expand the number of companies posting job listings on the platform, Deve has decided to switch to a freemium model. Interestingly, Deve plans expects Whitetruffle revenue to decline in 2013 because of this switch. But access to more data, he says, is much more valuable than short-term revenue.
"We're a data play," Deve says. "It's all about the data."
Getting off the Ground
Back in 2011, Whitetruffle was born out of a mutual frustration of Deve--who had been a candidate searching for a job--and Ali Behnam and Michael A. Morell, the co-founders of Riviera Partners, a top global tech recruiting firm headquartered in San Francisco. Morell and Behnam decided to incubate Deve and Whitetruffle within Riviera. Working side-by-side (in the same office, even) the three co-founders began building out a product that would address the concerns of both parties. Since its launch, Whitetruffle has raised $1.6 million from several investors, including Jerry Yang of AME Cloud Ventures (and co-founder of Yahoo), Scott Banister, Morado Ventures, and Kima Ventures.
"This industry--recruiting--is extremely broken on a number of levels," say Behnam, who is also an investor in Whitetruffle. "It's really a data-sensitive problem and we thought that we could create a better experience with technology."
This isn't the first start-up trying to leverage the cloud to help companies recruit more effectively.
Gild.com, for instance, scrapes the Web looking for lines of code written particularly well--and the developers who wrote them. Another, Talentbin.com, searches the entire Web--using open intelligence data from Facebook, Twitter, LinkedIn, Google Plus--to mine for potential applicants.
"The problem I have with those services is that they scrape the Web," Deve says. "Developers have no idea that they are on there. I'm not talking about Gild specifically. [Whitetruffle] is offering recruiters access to data about these people [that they've supplied themselves]. People don't want to be spammed by anybody, especially in the tech industry."
Michael Morell adds, "A lot have tried [to fix the recruiting problem], few have succeeded. You have software guys that don't understand recruiting, and recruiting guys that don't understand software."
Ultimately, Deve recognizes that for Whitetruffle to success, it needs to grow. "To realize our potential," he says, "we need much more data than this. We need to grow much bigger.
"At the end of the day, we will be very succesfull if we will show you one person, and you will hire that one person. That, for us, would be success."