Unlocking big data
As the online marketplace grows, so does the demand for analysing and leveraging big data. Hirers and candidates want more insights, faster connections and the most relevant employment information. As a market leader, SEEK is in a unique position to meet this demand. Big data drives innovation and creates exciting opportunities to experiment and evolve. But analysing big data requires a talented team to see past the numbers and interpret the findings to create viable, inventive products. To tell us more, we brought together a panel of experts to discuss how big data is driving innovation and growth across the business.
Introducing SEEK’s data experts:
Jacqueline McCudden is a Product Manager in SEEK’s Search team
Eugene Svistunov is a Lead Data Scientist in SEEK’s Marketplace Analytics team
Andrew Young is a Search Enrichment and Taxonomy Lead in SEEK’s Structured Data team
Aimee Jellett manages the Outbound Communications activity in SEEK’s Marketing team
Describe your role at SEEK and the data you work with.
Jacqueline: I focus on the relevance of our algorithms for job search and talent search. The data I use is quite user-specific at the moment. We’re in the midst of trying to obtain more user-centred data because we can’t really track very much. We have search rules that filter a user's journey but we don’t actually have a lot of unique things.
Eugene: Our primary focus is working on recommendations, suggesting jobs you might be interested in because you might have clicked on something in the past or applied for something in the past. We’re really focused on finding the best matches for the best candidates. We rely on the Data Services team to help us with that. We try to make the best use of all the job ads we have. Last year search became a real area of focus.
Andrew: We do two things; we try to improve search relevance by adding structure to otherwise unstructured content on both sides of the marketplace. The other thing we do is provide browsable structured data for the site itself; so the hierarchies that you can click through and the filters that can be used to narrow down the search results. In the Structured Data team we’re trying to add some structure back in so we can do more intelligent search matches.
Aimee: Our focus is delivering the most relevant career related content to both actively seeking and monitoring candidates. Big data assists us in delivering content and insights that are tailored to the individual, so that candidates can make informed decisions about their careers.
How has SEEK evolved over time and has the approach to using data changed?
“I think we’re in that really exciting next phase of SEEK now, where we’re truly leveraging our data.”
Jacqueline: I know that the journey for Talent Search has been very difficult, because people come to SEEK for different reasons. So we’ve struggled [to find] a successful measurement of success. What SEEK is now working with universities on and what we’re focusing on is exploring more granular metrics to measure success.
Andrew: Once we started looking at our data structure and where it came from, [we realised] it was largely mimicking what existed in print. We had a classification structure designed specifically to mimic print and we were looking to bring people across from print to online, which we did successfully as part of the first phase of SEEK’s business. If you look at where we are now, we’re trying to take advantage of online so that we can have a multifaceted view of information, rather than a singular classification of data. Being online, you can slice and dice your way through data any which way you want, and you’ve got more and more data points to analyse. So we can make more and more recommendations, and we can offer people more and more different experiences in terms of search and navigation. I think we’re in that really exciting next phase of SEEK now, where we’re truly leveraging our data and the possibilities that online forms, and we’re breaking the ties off where we came from.
Aimee: If you look at the journey of SEEK over its 17 years, we played a key role in taking print classifieds and making them available online. Now if you look at the market and companies like realestate.com.au and Coles, what's happening is that the customer's expectation for personalised information is driving the need for big data. Customers are becoming accustomed to receiving communications such as tailored emails and being retargeted on different websites. This is the role of big data and we're in a really exciting space where our focus on big data needs to shift and evolve with the demands of our audience and the competitiveness of the market.
Jacqueline: And our priorities have changed. The people we’re dealing with now want access to huge amounts of data, and want to use it.
Do you think sensitivity still exists in the marketplace in regards to SEEK’s ownership of big data?
Jacqueline: I think, from a candidate perspective, there’s kind of an expectation that you’re going to be seeing relevant results, so I don’t think there’s too much sensitivity around it – at least from the research and testing that we’ve done. From a hirer’s perspective, [it’s] us encroaching on their space with Talent Search. But we could easily provide Talent Search without a huge amount of big data.
Aimee: When it comes to sensitivity around data, we look to our customers behaviour and preferences and work to provide different levels of privacy to suit the individual. If you've ever been privy to one of the meetings where we talk about privacy, it's amazing the level of detail we go into to protect the privacy of both our candidates and hirers.
Does data have a lifespan? How long is it before data becomes outdated?
Andrew: Yes, particularly candidate data. It starts degrading the minute they’ve entered it. That’s something we talk about a lot and we struggle with. Eugene’s team spends a lot of time trying to work out [when] to stop using data, because it becomes less reliable. It’s just the nature of the business. We’re trying really hard to increase our engagement level with candidates so that they come back more often and update more often, so that the data becomes fresher and fresher.
Jacqueline: The most recent piece of information about a candidate is usually the most valuable, but older data is still relevant from a trend perspective. In our algorithms we have what we call a freshness component, so it freshens the results. Trends are really valuable to us because we learn a lot about the lifecycle of candidates.
Does SEEK work with any external parties to draw insights from the data?
Aimee: We work with different market research companies. Recently we developed some insights that will assist us in segmenting the information, driven by big data, that we share with our audience.
Jacqueline: That’s the benefit we’ve got by being a market leader, at least in job search; we have a lot of insights. I’m pretty excited with the evolution and the support that we’re getting to drive product through big data.
Posted by Nick Fulton, a Customer Service Representative in our Customer Service team at SEEK