What is usability testing?

Usability Testing looks at the extent to which a website — or app, product, service, anything that has a user — can be used by specified users to achieve specified goals. The test looks at effectiveness, efficiency, and satisfaction.

To create a new, effective website design many components are critical. Of course, the visual design and the branding are very important. But user testing reveals how the audience thinks about the design, if the design (and other factors) provide obstacles for website visitors to find the information they need or complete certain tasks.

User testing asks these questions:

  • Can the users complete the tasks they wanted to?
  • How long does it take them to complete those tasks on the website design?
  • Are users satisfied with the experience of completing the tasks?
  • During the web design process, when should you leverage user testing?

Often the website design is tested which provides tremendous value. User testing can also be very effective, even more effective, in earlier stages of the website design process:

  • in the discovery phase: to determine aspirations of the audience
  • in the architecture phase: to determine how content should be organized and structured, to design workflows

The earlier you test, the easier it is to fix problems because design and development have not invested time yet. Making changes is easier and less costly in early stages. In later phases of the website design, it can be very complex and time-consuming to fix usability issues, depending on what they are. If you wait to test until the site is built and implemented, it can become nearly impossible to fix usability issues, but it might provide insight for a redesign in the future.

Early and frequent testing is ideal. It keeps the user in mind at all stages, avoids mistakes, and creates effective websites.

How many people should you test?

User testing doesn’t have to be done with many people. Even the feedback from one person can add incredible value. Testing with 5 people will give great results. Of course, the more people provide feedback, the more reliable and robust the results will be.

A general guideline is that 5 users will detect ~85% of problems. Read why >