Smarter query intent algorithms for even better matches
To understand what a candidate truly wants when searching for jobs, you’d need to ask each candidate what it is they are looking for. SEEK is visited over 25 million times each month, handling over a million searches daily. That’s a lot of search queries to punch through, so an automated approach needs to be taken. What makes it even more challenging is to infer the intent from very few words. We speak to Duncan Norman, SEEK’s Search Product Manager, to pick that search puzzle apart.
For several years now, we have embarked on a journey to understand candidates’ query intent with SEEK. Understanding the intent behind candidates’ searches (past or current) allows us to deliver more relevant job opportunities to them and greatly improve user satisfaction. This in turn delivers hirers more relevant candidates.
Our team of search experts are currently developing a model for the identification of the candidates’ intent using their past interaction with SEEK. Simply put – what candidates are saying (typing) and what they’re trying to look for, and matching those up. The more candidates search, the better and more relevant search results we'll get.
You want the right candidates, not thousands of applications. Our search team are also working closely with SEEK's analytics team to write algorithms that take search queries and turn them into answers. These algorithms rely on many signals that make it possible for us to guess what the candidate might really be looking for. These signals include things like job ad content, location, salary offered and the candidate’s professional experience. This is an evolving piece of work and we’re continuing to develop these search algorithms and analytics to deliver you the very best candidate matches possible. Searching on SEEK is getting even smarter and more seamless.