Usability Analysis of a Federal Research Agency Website
A research library housed within a large federal agency wanted to move its website from a sharepoint environment to a Drupal content management system. My UX team was tasked with producing a comprehensive usability analysis of the existing website to assess the website’s design and determine what changes were needed during the migration.
I led the effort to evaluate the existing website. Our team included another UX designer and a project manager.
Our team established the following goals:
- Determine the usability of the current library website;
- Identify any strengths and weaknesses in the current website’s information architecture; and
- Develop recommendations for improving the website.
We encountered several challenges during the course of our work. For example:
- We were unfamiliar with the domain language and terminology, which was scientific and technical in nature because the library website is primarily used by scientists and researchers. Therefore, when assessing the information architecture and labeling of pages, we were often unsure whether the terminology was correct.
- We only had a few weeks to complete the assessment, a timeframe that coincided with a holiday season during which many of the website’s users were on vacation and therefore unable to participate in our research.
- The library team also wanted to use the redesign as an opportunity to highlight other services provided by the library that were not being significantly utilized.
We conducted a heuristic evaluation, card sorting, and user interviews. We took this multi-pronged approach in order to obtain the most detailed information possible despite the challenges noted above.
We utilized Jakob Nielsen’s heuristic evaluation principles, as those seemed to be the most appropriate for our needs.
Jakob Nielsen Usability Heuristics
We assessed the current library website using the above list, including details to support their findings. We then concurrently discussed each finding, evaluated the severeness of each finding (using the scale below), and prioritized the most serious issues to address. Finally, we identified other areas for which additional testing was necessary.
We used card sorting to evaluate the library website’s information architecture and label use. Because our work took place during the holiday season when many employees were on leave, it was not possible to conduct in-person card sorting. Therefore, we used online card sorting software (OptimalSort; see image below). We opted to use an open card sort method in which users organized select website pages into groups, and then labeled those groups. We solicited participation via the agency’s internal email listserve.
Web Based Card Sorting
While we had some data to show how users navigate the website, we still wanted to interview users and observe their behavior while using the site. Therefore, we recruited a handful of website users via referrals from library staff. These website users were asked to describe their job function and how they used the website for research.
Our primary finding was that most users were using the library website to get to PubMed (a search engine for life sciences and biomedical topics), even though the library website provided a PubMed search option within the website (see image below). Users were not aware of this internal search option because it was hidden under the marketing section. In addition, the PubMed option was not the default search option within this section of the website. Our assumption was that banner blindness and an overcrowded home page were causing users to ignore this area of the website.