Best Practices for Usability Testing In UX Design

Creative Cloud

Here’s the situation: you’ve created what you think is a fabulous product. You’ve invested tons of money in designing and building it, as well as marketing and selling it, but you’re not getting the results you had hoped for. Did you neglect to check if there were any usability issues?

If a user is unable to complete a desired task, then they’re not going to stick around and your product is not going to be successful—no matter how good it looks.

“These days it’s pretty easy to make new products—there are a lot of frameworks and tools out there—but it’s still really hard to make a great product, a product that people can understand and use, a product that makes people feel good,” said Aaron Walter, VP of R&D at MailChimp and author of the book Designing For Emotion.

“That’s so hard to come by and it’s definitely a competitive advantage if you can do it,” he added.

This is where usability testing (or user testing, as it is often called) comes into play.

What is usability testing?

Usability testing involves testing and monitoring user behavior as the user attempts to work through and complete a desired task. It’s a small step that is not to be overlooked and can save you a lot of time, money and embarrassment in the long run.


While there are a number of factors that go into usability testing, it can be broken down into a few major steps:

  1. Identify what needs to be tested and why (a new product, feature, etc.)
  2. Identify the correct audience (your desired customer)
  3. Create a list of tasks that the user will have to go through
  4. Recruit the right test users
  5. Involve the right stakeholders
  6. Apply what you’ve learned

Identifying the perfect recruit

Your ideal test user is the person who will actually be using your product.

“There are systems out there like, which is a great service, but you’re basically hiring people who aren’t necessarily your customers to pretend that they’re your customers,” Walter said. “Choosing someone who is actually your customer, who is actually likely to use this part of your product, is ideal so they’re not manufacturing their motivations.”

How many people should you test?

“You don’t need a huge sample size,” Walter said. “Somewhere between three and five will help you understand what the problems and patterns are.”

He points to Steve Krug, author of Don’t Make Me Think, who is also known for recommending small numbers because it’s what people have time for.

“Being able to fit this into your schedule is more important than having some huge sample size,” Walter said.

Design the test

Create the workflow that the test users will work through and make your objectives clear. You want to observe the range of errors or challenges that the user experiences.

Walter recommends asking them to use what’s called talk (or think) aloud protocol, encouraging the user to talk through their experiences and frustrations.

“When they’re speaking aloud you can get into their head a little more easily and understand where they’re getting confused,” he said. “It can help you understand things like maybe this language isn’t very clear here. Maybe the navigation isn’t organized very well. Maybe the primary call to action on the page is not as apparent as we thought it was.”

Recording the process through an app like Silverback can also give you something to refer back to and identify points where the users struggled.

Who should be involved?

Anyone that plays a role in how fast and how well that problem is addressed should be involved. These stakeholders could include the executive team, major decision makers, agency representatives and lead developers and designers.

Why the design team, or a representative of the design team, needs to be involved in usability testing

“You want the people that are designing and building the software to squirm in their seat as they watch the customer struggle and get confused because if they feel that pain and discomfort, they’re going to run straight back to their desk to fix it,” Walter said.

If someone is merely relaying information back to the design team, a lot can get lost in translation.

“It’s best to spend a few bucks, buy some lunch and invite those stakeholders to come eat and watch these usability tests happen live, or watch the videos—that often can work really well too.”

Conduct a quick post mortem

A brief chat with the test users after can sometimes expose a few extra things, but primarily the bulk of the learning happens by watching the behavior.

A few more things to keep in mind

Be sure to conduct extensive QA testing before doing any usability testing or else you could end up wasting a bunch of people’s time, including your own.

It’s also important to recognize the role human emotions play in testing, using it to both acknowledge your users and fuel your problem solving.

“You watch customers that get super frustrated and lost and they start to feel dumb, which is why it’s really important to tell them at the beginning that we’re testing the software, not you. They feel like there is a spotlight on them and if they can’t figure something out then embarrassment is heightened, frustration is heightened,” Walter said. “Generally I see that as a good thing especially if you have stakeholders watching that usability test.”

And finally, don’t lose sight of what usability testing is all about.

“A lot of assumptions are made in the design process of how people will use it, but there’s no hypothesis of where people are going to fall apart,” Walter said. “[Usability testing] is mainly just getting a first look at how people actually use this thing you’ve designed in the real world.”

The whole point is to ensure you’re releasing the best product possible into the world, one that not only meets the objective, but is also something you as a designer can be proud of.