CAST Usability Testing

“It was difficult to sift through all of the information and complete the tasks. I had a hard time figuring out where the information was housed.”

Synopsis: We tested seven users and found the biggest issue was the inability of users to find the information they were looking for. The final recommendation included an overhaul of the site information architecture, which is currently ongoing. 

I'm not going to lie, sometimes the truth hurts. Usability testing has the ability to uncover said truth - yet often this comes as a surprise, or worse, it can be met with downright outrage. As a researcher it is my job to provide clients with the appropriate data to move their product in the right direction. The purpose of this project was to run a remote usability test on CAST, which is an online tool developed to run scenarios to reduce nitrogen, phosphorous and sediment in the watershed to support environmental planning. 

To start, the test was sent to a total of 80 registered users of this tool. Unfortunately, only three users responded and completed this test, which forced us to use a panel of testers recruited from to round things out. While we had our suspicions on why we had such a low turnout there was nothing left to do but forge ahead. 

The results of the usability test were boiled down to two key elements: 
1. Simplify the login and navigational elements
2. Develop a clear information architecture

Tester Commentary:

“I wasn’t sure when to use the top ribbon and when to use the buttons lower on the page. I would also separate active functions like scenarios and reports from background and tutorial information.”
“I was unable to find several things, mainly because it was hard to search for stuff based on a location (e.g., BMP planning information and data related to Fairfax). I would add a navigation section that included the different locales/locations CAST is associated with.”

We found the majority of the issues came from testers unable to locate the information they were looking for. Since CAST is an incredibly powerful tool it is critical to organize it in such a way that allows users to accomplish their goals through the path of least resistance. This was enough of a red flag for us to suggest continuing with our research by conducting a tree test and card sort using Optimal Workshop. 

Depending on the results of these tests we could use the following frameworks to organize CAST: 

Topic: Under a topic based system, content on a website is organized by subject. Think: bookstore, i.e. biographies, science fiction, romance. The structure of CAST appears to follow topic-based organization; however if it were to continue in this manner it would benefit from a more focused topic structure as many users had difficulty successfully completing tasks during the user test.

Sequential: Content in this way is organized in a step-by-step format. Using a retail website for example, information is organized in such a way that a person navigates the website with an end goal of making a purchase. In this manner, one would need to complete a task before moving onto the next. An end goal for a CAST user might be to determine the most cost effective BMPs and the content would follow a step one, step two, etc. format. 

Audience: Using this scheme, content is organized for separate groups of users. For example, user groupings on an education-based website might be "students" and "teachers." Similarly, CAST groupings might include: Non-profits, Federal and Academic. This scheme allows for audience-specific content if there are large enough differentiators between the groupings.

Additional results of the usability test will also lead to an updated homepage with a streamlined login page and consistency with the navigational menu. A proposed flow would include three actions a user could take when arriving on the homepage. Also included in the flow is an integrated search and areas to view the documentation and download applicable source data. 


Although usability testing can easily go awry with a limited number of respondents, there is still an understanding of "some is better than none." As this project continues, we hope to run another test after the recommendations are finalized on the site.