Expanding data analytics capabilities help Seek personalise job search
- 17 June, 2014 17:56
Using data analytics to deliver personalised and proactive recommendations for job seekers has helped Seek to lift job application rates by up to four times.
The Australian online employment site launched a Weekly Round-Up email in February to provide candidates with highly targeted job recommendations based on their previous job seeking behaviour data.
To date, the email has been sent to 850,000 candidates and recommended more than 25 million relevant jobs. Candidates can see the number of job applications coming from similar candidates, the number of job ads posted for similar roles, as well as the top three employers of the week.
According to Seek, average email open rates are over 40 per cent, with more than 50 per cent of recipients coming via mobile devices. The service is available to all subscribers who choose to opt into the service.
Head of Marketplace Analytics, Antony Ugoni, told CMO the email has also lifted the likelihood of a candidate applying for a role by up to four times, based on comparisons with control groups who have not received the same data-driven recommendations.
The analytics capability and smarts behind the Weekly Round-Up email were developed by Seek’s Marketplace Analytics team, which launched 12 months ago under the group’s strategy division. The analytics team is tasked with looking at ways the organisation can better utilise big data and analytics to match jobs with candidates in a more robust, efficient way.
Today, the group sits at seven staff and has been supported by recent investments into database and software technologies.
“The priorities for us are a mix of what’s most important to the business, balanced against the bits we can chew first,” Ugoni said. “We have roadmap in place that is going to take us to a point where we’re taking full advantage of the data assets available to us, but there are small steps to get there.
“We’re analysing job ads really well at the moment; with CVs, we’re a little bit behind that, so it’s a a balancing act trying to make sure you have an eye on strategy but also delivering value in the meantime.”
Ugoni said the first thing the analytics teams needed to do was spend time understanding the business’ data assets and what’s available, as well as the latency and accuracy of the data sources. But to be successful, he stressed the importance of articulating the business problem before utilising data.
“You only derive value when you understand the business problem,” he claimed. “We spent as much time working with the business and strategy team to understand the business problems first, then we point back to the data and use our analytics skills to solve that problem. We never start with a data point or algorithm, it’s always about the business.
“The good thing for us is Seek has clarity on what the key business problems are, so the rest is relatively easy.”
In the case of the Weekly Round-Up email, Seek had identified a need to ensure good candidates who are not necessarily active at the time a great job becomes available are still made aware of relevant positions.
“The hirers had told us they want to get to access to as many good candidates as possible, not just the active ones,” Ugoni said. “The marketplace needs to be more efficient; we need more candidates in the pool.
“Seek understands its position in the market, has a unique position in terms of some of the data assets we can pull together both on the candidate and employer side, and we see an opportunity to make the jobs market a more efficient place.
“These innovations make it better for everyone.”
Ugoni said he’s now looking to optimise the algorithm behind the Weekly Round-Up email service to further improve the service.
“A big focus for us is keeping an eye on current algorithms to make sure they’re not deteriorating, and to find the pockets in the candidate landscape where we could do better, as well as find areas where the algorithms are doing really well to understand why they’re doing well and see if we can bring that learning to the rest of the hiring market,” he said.
The Marketplace Analytics team will also launch several complementary algorithms shortly, which Seek’s product team can then take into the field, Ugoni said.
While his team is largely relying on internal data sources, Ugoni is increasingly working with the strategy teams to look at relevant external data sources that can help broaden its data-driven services.
“This will help us hone the products we are building, and that external macro level data is being taken advantage of throughout the organisation,” he added.
More on how companies are utilising data analytics