Case Study
A lightweight, mixed-methods usability audit diagnosing friction on a crucial public-access academic platform.
123Origin
I noticed usability friction on SciELO South Africa and ran a self-initiated audit to diagnose it. SciELO South Africa is a free, public-access academic journal platform that plays a crucial role in South African scholarship.
Timeline
2 weeks
Role
UX Researcher
Platform
Academic Web Platform
Investigate
Examine where users encounter friction while completing common research tasks.
Uncover
Identify patterns of usability issues across novice and expert academic users.
Propose
Recommend clear, evidence-based improvements grounded in usability heuristics.
2
Participants
17
SURVEY RESPONSES
3
RESEARCH METHODS
3 core
Findings
Approach
I evaluated SciELO from three complementary perspectives: the interface, individual researcher experiences, and broader user feedback. This allowed me to identify usability issues, understand how they affected real research tasks, and validate recurring patterns.
01 — Heuristic Evaluation
I started with a heuristic evaluation to identify potential usability issues in the interface before testing how users experienced them.
I evaluated SciELO against Nielsen’s 10 usability heuristics, focusing on core research activities including search, filtering, navigation, error recovery, and information scanning.
Visibility of System Status
No loading indicators or actionable feedback during longer processes such as filtering or downloading PDFs.
User Control & Freedom
No clear way to reset filters, while limited mobile responsiveness restricted flexibility.
Aesthetic & Minimalist Design
Heavy text density, limited whitespace, and outdated iconography increased cognitive load and made results harder to scan.
02 — Usability Interviews
The heuristic evaluation identified potential usability problems, but I needed to understand how researchers actually experienced them during real tasks.
I selected two participants with complementary perspectives: a graduate student representing a typical researcher, and a metadata librarian with expertise in categorising scholarly information and conducting research.
I used semi-structured interviews and task-based observation to explore how participants searched, navigated, filtered, and interacted with academic content.
The sessions revealed behavioural workarounds and points of friction that weren't apparent from the heuristic evaluation alone.
03 — Validation Survey
The interviews provided depth, but with only two participants I needed to determine whether the issues identified were experienced more broadly.
I surveyed a broader group of researchers about their experience with SciELO’s core features.
The responses helped validate the frequency of recurring issues identified through the evaluation and interviews.
Findings
Across the heuristic evaluation, interviews, and survey responses, several usability issues consistently surfaced. These findings connected interface-level problems with the behaviours and frustrations experienced by researchers.
Finding 01
The heuristic evaluation suggested search visibility issues, which user testing confirmed:
"I couldn't immediately locate where to start searching. The homepage didn't show a clear search box or icon."
"I wasn't sure if there was a search function at all. It wasn't obvious, which made me hesitant."
Survey Validation
68%
of survey respondents struggled to locate the search bar.
Finding 02
"The filters felt buried. It took extra clicks to find them, disrupting my workflow. … a missed opportunity for learning."
"The filters were confusing. The labels didn't explain what they meant, so I wasn't sure which ones to pick."
Survey Validation
74%
of users found filters difficult to understand.
52%
avoided filters due to uncertainty on how to use them.
Finding 03
Survey Validation
60%
of users found navigating categories confusing or requiring too many steps.
"I often used my browser's search instead of the site's navigation because it was easier."
"I got lost a few times — I didn't know where to click to find related papers or get back."
01 / No Clear Path In — Users land on a rigid index page with no visible route back to browsing or the homepage, forcing them to start over rather than navigate forward.
02 / Dead-End Selection — This index-selection screen offers no breadcrumb or 'back' option, so users lose their place in the site the moment they commit to a term.
03 / No Way Back or Forward — The citation page buries any link to related work or the original collection, leaving users stranded with no next step but the browser's back button.
Strategy
01
Introduce a prominent, persistent search bar in the primary header with clear iconography and placeholder guidance.
02
Rewrite filter labels in user-facing language, group by intent, and expose the most-used filters by default.
03
Restructure the Information Architecture around research tasks (browse → search → read → cite) rather than internal content categories.
04
Add lightweight in-context tooltips to introduce advanced search operators and filter behaviours on new users.
Reflection
This project let me combine two passions: research and improving access to education, while deepening my qualitative research practice.
My background in sociology shaped how I approached the interviews by asking thoughtful, people-centered questions. This was a look at how design can either support or block access to knowledge, especially for users who rely on public tools like SciELO.
After interviewing Zizipho and Ayanda, feedback from my mentor reshaped how I think about usability testing. I learned to set more realistic scenarios, consider users' real context (like when and how they last used academic tools), and avoid guiding participants too much during tasks: changes that make space for more honest, useful insights. Even with just two interviews and a small survey, this approach surfaced real, actionable findings.
Follow up with low-fidelity prototypes to test possible solutions.
Run a comparative analysis against similar academic platforms like Google Scholar or JSTOR.
Expand the participant pool across disciplines and institutions to strengthen the evidence base.
Share findings directly with stakeholders at SciELO or in the academic community.
