SEEK quadruples click-through rate with user analytics

By on
SEEK quadruples click-through rate with user analytics

Job applications increase six-fold.

Job search company SEEK has increased user click-through rates four-fold and sextupled the number of actual job applications thanks to a text-mining and analytics research project with RMIT and the University of Melbourne.

The three-year "Seeksy" project incorporated a query amendment system into SEEK’s search function and made it possible to expand search terms to include uncommon or colloquial variants.

The expanded array of variables comes from text mining techniques informed by analysis of user history and behaviours.

Seeksy can therefore cope with synonyms for the same job - think  “governess” and “nanny” - that describe the same job but are used only in different places or groups.

It can also modify overly-specific search terms to broaden the range of results presented to users to offer a more comprehensive, targeted experience.

A/B testing of the technique showed that click-through rates with the Seeksy tool increased four-fold for a select group of the most problematic search fields, while the number of job applications submitted increased six-fold.

Project Lead at RMIT associate professor Lawrence Cavedon said analysing and improving job queries has been vastly different to traditional web searches.

“Job search involves different behaviours and goals to standard web searching, so we started this project by exploring user behaviours on the site through log data analysis,” Cavedon said.

“This showed us where to focus on designing automated techniques that would really help people find what they were looking for more easily.”

He added that it was satisfying to know the improvements were based on evidence of what users actually need to get better outcomes from the service.

The project, now in its final phase, kicked off in 2015 thanks to a $394,000 linkage grant from the Australian Research Council.

Got a news tip for our journalists? Share it with us anonymously here.
Copyright © . All rights reserved.

Most Read Articles

Log In

  |  Forgot your password?