One of the foundational elements of user experience (UX) design as a discipline is usability. While your marketing team or website designers might have ideas about information architecture best practices, those will only take you so far if users aren’t connecting with the organizational structure or language you create.
As you develop your information architecture (IA), you want to start by hypothesizing what you think will work best. But before confirming or finalizing anything, make sure you get input from end-users to ensure your IA works as intended.
A usability test allows you to evaluate how effectively your IA meets your users’ needs. According to usability.gov, the goal of running one is to “identify any usability problems, collect qualitative and quantitative data and determine the participant’s satisfaction with the product.”
On top of helping you ensure a positive user experience, usability tests also help you evaluate accessibility. Part of the test should include ensuring that users understand what they’re supposed to do with your user interface and are able to take the desired actions.
How to Run a Usability Test for Your Information Architecture
1. Identify the problem
Like most experiments following the scientific method, the first step of running a usability test is defining the problem you’re trying to solve.
For IA, the problem is almost always that users aren’t able to find the information they’re looking for quickly and efficiently.
2. Determine opportunities for improvement
After defining the problem you’re trying to solve, the next step is to come up with potential solutions. Start by ideating alternate categories for your main navigation and subpages.
To help inform this process, you can do competitive research to see how other companies in your industry organize their website content. You can also do keyword research to gain a better understanding of the terms users are looking for. Finally, you can conduct some user research, like sending out surveys or a card sorting exercise.
3. Test potential solutions with users
After conducting research and building out a new navigation and site map, then it’s time to see how users engage with your new proposed IA.
Typically, you do this by asking a user to complete a specific task. For example, if we were running a usability test on New Breed’s site, we could take the user to our homepage and then ask them to find information about our revenue operations product. That page isn’t accessible directly through our main navigation, so we’d be examining the way the user goes about looking for it.
Your test facilitator will observe the user trying to complete the task, asking them questions about what they’re doing along the way. In turn, participants will provide feedback and explain why they’re choosing to take each action.
As you’re running your usability test, you should be collecting both qualitative and quantitative data, such as:
- How much time does it take for them to find the content they’re looking for?
- What are they emoting: frustration, confusion, ease?
- What kind of information are they telling you as they narrate their thought process?
To get accurate results, you need to be consistent in how you give instructions and record user behavior.
To understand the effectiveness of your new IA, you need baseline data from your existing system to see if your new one is an improvement. So if you don’t have that benchmark already, you’ll need to collect it during your usability test to establish a control.
Test group makeup
Your test group participants should be the type of people you’ve built the website for. Especially with B2B or SaaS websites, you’re typically marketing to people with some knowledge of what your industry is or what your solution does.
Because you have this niche audience, it’s okay if the language you use in your IA won’t resonate with someone completely outside of your target audience. So, you can narrow down your usability test participants to just the type of people you want navigating your website.
However, it is still important to test your IA on people with varying degrees of background knowledge. Your participants should range from people who are in your industry but unfamiliar with your company to current customers that know your offerings intimately. This way, regardless of what stage in the buyer’s journey a user is at, they’ll still be able to have a positive experience on your website.
You should always have a test plan built out before starting. Vet that plan internally before working with external users to ensure that your test will actually help you reach your desired objective.
If you’re trying to get the user to go to a specific page, you want to make sure that the question you ask to prompt the action is clear enough that the user can complete the test, but not so clear that it’s leading.
Your plan should also outline how you’ll be measuring the results. Define your KPIs and identify what data you’ll need in order to know making this proposed change will be beneficial to users.
Kelly is the Manager of Web Strategy at New Breed.
Connect with the author