WavebreakmediaMicro - Fotolia

Manage Learn to apply best practices and optimize your operations.

What usability testing methods can help me get started?

Usability testing is about solving existing issues on a given interface, but how do you detect those problems? Expert Matt Heusser offers his insights.

When thinking about the publicly available information on usability testing, my mind thinks of two things: mystery and expense. A quick Google search bears this out; I tend to find brief overviews with vague mentions of a "lab" environment and a lot of consultants selling training courses.

Fortunately, it does not have to be that hard, nor that expensive. If you have a budget, you could set up a series of rooms with one-way mirrors equipped with video recording tools to document both the user's expressions and the screen at the same time. Then we find a way to bring in, schedule and compensate our users for a few weeks while they try to use the software.

With that said, not everyone has an extensive amount of time to set up a lab and conduct interviews and analysis, complemented with a budget measured in hundreds of thousands of dollars. Instead, we might want to start with something small.

Two steps to consider regarding usability testing methods

Before anything else, usability testing is having a participant using the software in something close to real-use conditions. The elements are designed to break the groupthink phenomenon. When we build and test a user interface, we tend to understand it completely. In the book Made to Stick, authors Dan and Chip Heath call this the "curse of knowledge" -- that we can't imagine a world in which the other person does not understand the process.

For the first step, you will need candidates. As you find participants, decide what their process will be. For example, an accounting package might require registration, entering company information, setting up bank accounts and receiving payments. In this scenario, a realistic user is most ideal -- someone with enough booking or accounting knowledge to be able to use the software without a user guide. I'd suggest typing up these use cases, in some detail, and having them reviewed by a friend. Then, when it comes time to test, you can pass the use cases to the worker and have confidence that the confusion they have is about using the software, not about the task.

For a website, the participants might be answering questions that include, "Can they find a contact phone number and email?" or "Can they find common questions about our services?" You might get these questions from existing customers.

The second step is to brief the participant. Inform them your job is to remove obstacles from adoption and then find out when they are confused. Give the user the set of questions and watch what they do. If you see a confused expression; if the mouse floats around without clicking for an extended period of time; or, especially, if they say something or ask for help -- these are things to notice.

In Oleksandr Berchenko's article on software usability, he categorizes usability testing into a pyramid with utility at the bottom. Utility asks if the software is capable of doing what we ask it to do; testers can test for this without outside help. Berchenko splits usability into adoptability (is it an easy switch from what people know before? is it familiar?), learnability (can we pick it up quickly?) and efficiency (can I do powerful things quickly?). Berchenko puts identity at the top of the pyramid, where a user associates their sense of self with the application -- creating users who are unlikely to switch and even evangelists for the product.

You likely won't get that from a list of tasks and a user you recruited in the hallway, but you could get started tomorrow, and that's probably enough for now.

Usability testing is about finding problems in an existing user interface. If you find problems, it means rework. Walking users through prototypes or designing for user experience can help detect any issues earlier. Ultimately, you probably want both. Usability testing is relatively easy to add at the end this time, on this project that's 80% done already. It will be easier and cheaper to do if we design the experience first -- but that is a different article.

Next Steps

The difference between usability testing vs. user acceptance testing

How to bring UX Testing and software testing closer together

CTO says RSA aims for usability

Dig Deeper on User acceptance testing

Join the conversation


Send me notifications when other members comment.

Please create a username to comment.

What steps do you take to ensure accurate usability research?
A corner stone would be knowing your users. Design personas as an approach seems to be gaining more popularity, and they go into very specific details telling about users' attention, interests, and operational habits.
One thing we try to do is have the UX people work with different users on the design, rather than using the same user or the product owner every time. This helps overcome the “curse of knowledge” that Matt mentioned.
I have two thoughts.  One is embedding some kind of video and voice recording into the test so you can capture the screen along with it.  Another method might use question surveys to see where a user would expect to find the information they need for a particular test, perhaps with mock-ups.

Or you could use a tool like GUTAR-UX www.guitarux.com  Which provides a mechanism where you can do all of that in one experience and gather some empirical data. 
The article indeed suggests some good starting points.
I'd like to recommend UseIt.Com (nngroup.com) as another excellent source. These usability heuristics (https://www.nngroup.com/articles/ten-usability-heuristics/) are also portable as testing heuristics.
Things I consider with usability testing.

  • Flow of a procedure (does the place I start to where I end make sense?)
  • Intuitiveness of the interface (can I find my way from Screen A to where I need to perform the next operation on Screen B?)
  • Information that's available.  Is it organized in a manner that makes sense, is the information actually valuable, does it fit on the screen?
  • Fit for use. Does it make sense to use for the use case we are trying to achieve?