Recruitment firm Hays has increased its job search to application conversion rate by 10 percent after deploying the Google Cloud Talent product to better match candidates to roles.
The deployment applies machine learning algorithms to job descriptions, searches and profiles that evaluate the intent of jobseekers and the content of different roles.
It then automatically matches individual candidates to the jobs they want, while also enabling employers to quickly find the most appropriate applicants for their business.
Visitors to the Hays website are then provided with a personalised view of roles directly related to their skills and needs. They also show users other jobs that the candidate matching algorithms predicts may be of interest, such as presenting applicants searching for project manager roles with additional adverts for programme managers.
This system is based on an ontology specific to the recruitment sector, which allows it to connect relevant jobs to candidates. This can be a complicated task for a business that operates in 33 countries, each with their own vocabulary for professions. For example, someone looking for a "server" in the UK would expect to find technical positions, whereas in the US, they would want waiting staff.
Hays had weekly calls with the Google development team to ensure the product could be deployed at scale around the globe. This approach allowed them to deploy the system in every country where Hays operates without requiring the time-consuming build of regionalised infrastructure.
“The things that we were particularly pleased about was the ability to work collaboratively," Hays CIO Steve Weston tells Computerworld. "Because you wouldn't necessarily associate that with a number of vendors.”
The recruitment firm uploaded all its jobs into the Google Cloud service, where Google would associate it with the appropriate ontology.
Hays then evaluated the success of the project by evaluating the results of different stages of the application journey, from applications to shortlist and interview and finally the offer.
In the UK, the FTSE 250 company says it received an additional 170,000 monthly searches in the first quarter following the move to Google Cloud. Surveys showed that 60 percent of applicants believed their search results as highly relevant to their professional backgrounds.
Personalised enagagement
Weston admits that Hays was initially reluctant to make the move. The business, which celebrated 50 years as a recruitment company last year, had found a legacy system that served it well, but Weston recognised that its job search tools could be refined.
After having "tried every vendor in the world", the company ensured that the Google Cloud Talent system would be an improvement through A/B testing and what Weston calls “red eye testing”: virtual trials with real users that helped identify where the algorithm could be refined to deliver better search results.
“My take is always that being agile is about speed and agility in a controlled way, and that's really important to us," says Weston. "If there's a little bit of caution, it's just about making sure.”
The move reflects a wider shift towards personalised engagement with job seekers. Hays has tried to replace the traditional recruitment method Weston calls “post and apply” with a new strategy the company has dubbed “find and engage.”
Hays hopes this will help the company to retain clients for the long-term by tailoring the job search experience to the individual needs of both candidates and employers at the specific time of each listing and application.
Weston credits the success of the deployment to the close collaboration with Google and a willingness to adapt plans over time.
“You've got to iterate," he says. "Don't always assume the first thing you're given will be 100 percent and work. You have to have the approach that you're going to test and learn and work collaboratively in a very quick way to make it work.”