Ever since reading Steve Krug’s Don’t make me think a few years ago, I’ve wanted to run my own usability testing sessions. Usability testing is where you sit with people who have never used a site before, (preferably not web professionals) and discuss the site with them as they try and use it. Often their comments and mouse movements are recorded so you can go back and review what they said and did.
The idea is that when you are working on a site you can become so immersed in the project it is easy to forget what normal people actually want out of the website. Business objectives (e.g. “we must sell products and gather email addresses”) and technical constraints (e.g. “we can’t build that function and we have to display the data in this way”) often get a lot of priority on a project, and it’s the job of an IA to focus on the end user.
Usability testing is where you sit with people who have never used a site before, and discuss the site with them as they try and use it
Usability testing is a key tool in doing this – and importantly it provides evidence with which to inform your judgements and back up decisions. A video of representative users saying “I don’t want to give over my email address” or “I don’t understand the data when it is presented in that way” can be a powerful persuasive tool for making changes.
Another important point is that you can test sites when they’re in the prototyping stage – i.e. they don’t have to be fully built. You can create a clickable web mockup of full colour designs, or even wireframes and still get invaluable feedback from the process.
The bottom line for me is that it just seems such a simple and easy way to vastly improve your web sites. At my previous job I wrote a proposal offering to run low budget, ‘lost-our-lease’ usability testing sessions as advocated by Krug. The idea was to do cheap, frequent testing rather than expensive agency testing (which happened about once a year).
In the end they never went for it, and continued to outsource the work. But my new employer seems to be a bit keener to do usability testing, and let me loose to test the configurator for a large car brand we’re working on. A couple of weeks ago I ran my first ever usability testing sessions, with 6 volunteers from my work spending around an hour with each of them on the site.
The bottom line for me is that it just seems such a simple and easy way to vastly improve your web sites
The product
The website I tested was a car explorer and configurator system designed to give users a visual tour of various car models (the ‘Explore’ section), before allowing them to choose and configure their own car (the ‘Configure’ section). The ‘Explore’ section was already fully functioning and live on the web, while ‘Configure’ was in prototype form. I created a clickable version of the visual designs for users to interact with in the test.
In the Explore section you can view one of four models of luxury car in 3D, navigate around it, look inside at the dashboard and interior and change the colour and wheels with real-time updates. In ‘Configure’ you can pick your options from a list and see the price and car photo update in real time.
The outcome
The aim of the exercise was to get 4-6 key recommendations for improving the product, which could then be implemented to improve the user experience. We achieved this and made recommendations to the client, including video clips and justification for the points we made. Overall it was really good to see the users interacting with the system and I’m convinced we made genunie improvements to the system. I hope we can do it again on another project.
