Usability testing: more than meets the eye

By Hilary Stephenson | 11/11/2012

Every usability study is different from the last and the best approach depends on a number of factors, including budget, timescales, UI complexity and the intended audience. However, simply put, there is no better way of seeing if an interface meets its intended purpose than to test it with its intended audience.

Of course, as providers of web-related services our team is well aware of usability principles but even the best designers and usability professionals cannot possibly predict the usage patterns and behaviours of all users.

In all our testing activities, we take a task-based approach, asking users to navigate through our UI designs with specific, real-world objectives to achieve. It’s vital that the test participants represent the eventual site or application user audience, but recruitment for studies can be a major challenge. Although we genuinely believe that any task-based testing is better than none at all, we work with market research partners and our clients to identify a truly representative sample.

One of the techniques where user recruitment is vital is eye-tracking, a testing approach that captures real-time interactions with an interface, recording their eye movements and navigation steps as they work. We have been working with Acuity ETS, specialist providers of eye-tracking equipment, analysis software and training services in recent weeks. Eye-tracking provides:

  • Usage statistics and analytics
  • Gaze plotting to show the movement of our gaze around the interface
  • Heat maps showing how long users linger on particular areas
  • Comprehensive reporting

We conducted a series of eye-tracking sessions for Citizens Advice, which resulted in some immediate improvements and recommendations across  several of their web channels.   We also took the time to track our own site, which led to an interesting discussion with our design team!  Check out our videos on the WeAreSigma YouTube channel if you want to see some of the test output.