Day One: Usability Testing
Content & Activation Warning: Domestic and sexual violence
TL;DR
A local nonprofit was hoping to increase the usability on their website for folks who were in a moment of crisis. Through evaluative research and think-aloud usability testing, my team and I were able to do a mid-fidelity redesign of their home page to increase visual hierarchy, clarify language and boost organizational trust.
Team: Mackenzie Leach, Brian Spindler, Eric Rathke
Role: Researcher, Testing Moderator, UX / Product Designer
Methods: Evaluative Research / Usability Review, Moderated Remote Usability Testing, Think-Aloud Usability Testing, Affinity Mapping
The Problem
In 2020, many of us were asked to Stay-At-Home as a way to prevent the spread of Covid-19, but for victims of domestic or sexual violence, home was not safe. Day One, a 24/7 call center for victims of interpersonal violence in Minnesota, connects people to the support they need. Their team asked us to evaluate their website’s overall usability, specifically ensuring that it has clear hierarchy, find-ability, and accessibility for those who may be in crisis.
How might we quickly and accurately direct folks to the resources that are appropriate for them in their situation?
The Recommendation
While Day One is providing an incredible service to the entire state of Minnesota, there are several ways to make the process to getting help easier for website visitors. Three themes emerged:
Trust is imperative; introduce the organization right away and maintain trust by updating broken links
Visual hierarchy is unclear; don’t make website visitors think. Layout information where they need it, when they need it.
Words matter; avoid vague language and use visuals as much as possible
The Process
Understanding the Current Model
After receiving the project brief, the team individually completed a usability study using Neilson Norman Group’s 10 Usability Heuristics as a lens to analyze the website. We received site goals and primary tasks from the organization, so as I was exploring the main functions of the website, I was considering what how might we create scenarios to guide users through the site and identify a “happy path” to get to what they need.
Utilizing Trauma-Informed Practices
Based on our individual findings and organizational goals, we created a test plan and user scenario script to evaluate how well the site currently serves those goals, both for victims in danger of domestic violence and concerned family/friends/employers looking to get support for their loved ones. Because the subject matter was sensitive, careful consideration was made in writing the script to ensure participants felt comfortable and scenarios weren’t triggering. In addition, several layers of consent were put in place and trauma-informed communication methods were used to create a safe container for testing. Five participants were recruited and the team rotated between moderating and note-taking.
Making Sense of What We Heard
The team remotely analyzed the data using Zoom and Miro, grouping quotes and pain points to see the themes that arose. As with most group work, we diverged and converged several times to allow for all modes of thinking and communication to flourish.
I then took our rough synthesis and began even more sense-making, trying to identify the key themes based on what we heard. A few key things emerged.
Trust is Imperative
At first glance, participants were unclear exactly who One Day is and what services they provide. When considering how to help loved ones in need, some participants wanted to ensure One Day was a trustworthy organization. First impressions ranged from clinical mental health services to an informational website for domestic violence. Many needed to scroll and click before they understood the services provided. In addition, crisis hotline numbers on the top of the page were only noticed by 50% of our participants, hiding one of the key features from site.
Visual hierarchy is unclear
Similarly, while the “Chat Now” button is at the top of the page, three of the participants didn’t mention it until they’d been on the site for 15+ minutes. In addition, it’s a mental model mismatch. Fortunately, once identified, participants could understand its value.
Because the top navigation was organized by the type of violence instead of unique Calls-To-Action, each research participant chose a different route to find shelter and support services. The clearer CTAs were found on the footer, which 3 of 6 participants used to find shelter.
Words matter
In times of crisis, information overload and sorting through dense text is a heavy cognitive lift. Pages of heavy text can be difficult for people to process, especially those with accessibility needs. Several participants wanted information to be more visual, creating a lower barrier to understand the information.
Similarly, “Resources” was vague enough where participants didn’t know what to expect. Surprised by what they saw, we heard several opportunities to use more precise language and a sorting feature to minimize effort for future website visitors.
Based on the findings, patterns started to emerge regarding the pain points that research participants found most challenging. I broke down the recommendations into three categories:
Within each category, I prioritized most critical (high impact to website visitor / high impact to website goals) to least critical (low impact to website visitor / low impact to website goals), which can be found in the Findings and Recommendations Report.
In addition, I put together a mock prototype of what a new homepage might look like with some of the recommendations were applied.
The Next Steps
If I had the opportunity to continue on this project, I would recommend one more round of research that considered the following:
Broaden The Research Participant Pool: Testing for culturally appropriate services was difficult in this testing simply because of demographics. Conduct usability testing with clients that fit the demographics you hope to serve (immigrant populations, Spanish speaking, Somali, etc.)
Highlighter Testing: Several recommendations were related to content. Consider testing each page using “highlighter” methodology with Subject Matter Experts and potential website visitors to understand hierarchy of your content
Card Sort Testing: Identifying what language to use in the navigation bar could be identified using a card sort test. Similarly to the Highlighter testing above, testing with Subject Matter Experts and potential website visitors is recommended.