job hunting
jobandtalent

Jobandtalent Raises $3.3M For Its Linguistics Algorithm-Based Approach To Recruitment

Next Story

Samsung Joins Apple, Google And Microsoft With First Global Developer Conference October 27-29

U.K.- and Spain-based recruitment startup Jobandtalent, which uses linguistic analysis to alert candidates to jobs they might otherwise have missed, has just closed a €2.5 million ($3.3 million) funding round, led by Spanish investors including Kibo Ventures, Félix Ruiz (co-founder of Tuenti.com) and existing shareholder Alfonso Villanueva.

There are scores of startups attacking the laborious problem of job hunting, from the seeker’s point of view, and the difficulty of locating quality candidates, from the employer’s. Jobandtalent is (yet) another purporting to take the strain out of matching candidates to relevant opportunities and vice versa — but their (they say unique) twist is a linguistics-based algorithm that parses job ads and seeker CVs to locate relevant pairings, thereby allowing passive candidates to be more easily folded into the mix.

If its spiders return with a match between a candidate and vacancy jobandtalent automatically sends a message to the candidate to notify them about the role. Its system also ties in with social networks such as Facebook to foreground any relevant connections job seekers might already have in their social networks linking them to a prospective employer so they can request an introduction.

The linguistics-based approach is a relatively new twist for the startup — which pivoted last year — and it’s that pivot that has attracted this latest investment. Prior to pivoting, and since being founded in 2009, it had been testing different approaches to tackling recruitment issues, including career guides and a social media approach, says co-founder Juan Urdiales – raising €1.1 million in total angel funding at that time.

Its linguistics algorithm, developed in collaboration with HR and big data experts, with input from PhDs from the IULA (Research Institute for Applied Linguistics) of Universitat Pompeu Fabra in Barcelona, identifies and recognizes linguistic patterns within the structure and phrasing of  job adverts and CVs, and then converts them into data points to match candidates to suitable jobs. Jobandtalent claims its algorithm can create matches “even if they do not match a candidate’s or recruiter’s specifications word for word” — so it’s presumably doing something more sophisticated than keyword matching.

“We do big data analysis,” says Urdiales, explaining that this includes linguistic analysis between candidates’ profiles, and the job postings in its database. “We don’t only match the title of the position and the title of the job posting. We match everything. All the description of the profile — not only the title, but what the candidate has been doing when he was in that position and also we match… all the description of the job,” he says. It does also currently employ three internal recruiters (aka humans) to fine-tune its algorithm — to “sanity check” matches. But the ultimate aim is to have the algorithm taking all the strain.

Urdiales says the extent of the linguistic analysis of its algorithm sets jobsandtalent apart from its closest rival recruitment startup Bright.com which does semantic matching but apparently mostly parses just job titles and positions. “Our approach is much more advanced, uses much more information and is much more accurate,” Urdiales adds.

He pegs the current accuracy of the algorithm at 50% of matches being “totally accurate” (up from 20% when the beta algorithm was released about a year ago) and the other 50% having “a certain level of accuracy — but we cannot consider a perfect match”. The ambition, he says, is to get the algorithm as close to 100% as possible. “We will be happy if we can achieve 80% of accuracy,” he adds.

The core idea behind using an algorithm to parse job ads and profiles is that it saves times — and allows hundreds of thousands of jobs from different recruitment boards and sites to be speedily crawled. The method also foregrounds jobs and candidates that might otherwise miss each other, and therefore potentially increases the number of passive candidates alerted to job opportunities vs active job seekers. Passive candidates are generally valued highly by prospective employers since they may be the best suited to the particular role needing to be filled.  

Since pivoting last year jobandtalent says it has signed up more than 200 “leading European and global companies” as employer customers — including Accenture, Morgan Stanley, Ericsson, Booking, L’Oréal and Deloitte. It also now has more than 500,000 registered users, and employs 20 members of staff itself, working mainly in technology, big data and marketing.

Urdiales said the new funding will be used to improve the accuracy of the matching algorithm and also to develop native mobile apps. It also plans to focus on boosting user-adoption in the U.K., and prepare its platform for launching into new markets — with a focus on the U.S. and Latin America. It’s aiming to launch in the U.S. in Q1 2014, according to Urdiales.