Richmond Library IA Evaluation
Can students find what they need without learning the library’s language first?
Role
UX Researcher
Timeline
1 month · Spring 2026
Team
4 researchers
We usability-tested the current site with 8 students to learn why they struggled, then tree-tested the proposed IA with 7 more on the same five journeys.
Context
Boatwright Library partnered with Pratt’s Digital Experience Center to evaluate its website navigation. Our job was evaluation, not redesign: test the current site, test the proposed IA, and show where each succeeded or failed.
The current site, with OneSearch front and center
Method
research discovery
borrowing & delivery
study spaces
events & programs
wayfinding
Success: reached the right place. Direct: got there without backtracking. First click: started in the right section.
Findings
Browsing felt unpredictable. Students were confident with search but unsure when browsing the navigation.
Policies instead of actions. Students looked for “request” or “delivery” and found policy pages.
Labels to decode first. Room booking existed, but students had to choose between “study zones,” “study rooms” and “facilities.”
first-click accuracy for room booking
overall task success on the proposed IA
original pain points still unsolved
Success by journey, proposed IA (n = 7)
Room reservation
71%
57%
Book request
43%
14%
Peple Lecture
29%
0%
Research guides
14%
0%
Music library map
14%
0%
Success
Direct success
Physical spaces were the clear win. Academic tasks all stayed below 50%, and three had no direct successes at all.
Before
Facilities & Classrooms: read as classroom-only, then a choice between study zones, study rooms and facilities
After
Libraries & Spaces: one recognizable home for every physical-space task
First click correct
100%
Reached the right place
71%
Got there directly
57%
The label got everyone started. The pages after it are where the remaining work is.
Book request (43%): split between “Find & Borrow” and “Help & Support,” so students guessed. Peple Lecture (29%): still filed under the wrong section. Research guides (14%): students defaulted to OneSearch; how guides relate to it stayed unclear. Music library map (14%): still reachable through only one path.
Recommendations
Impact
High
Medium
Low
Each recommendation was presented with before/after mockups
Outcome
We presented our findings to the library team in May 2026. The Digital Scholarship Librarian called the feedback helpful.
8 and 7 participants give direction, not proof, and we only tested students, not faculty or staff. The natural next step is a second tree test once the recommendations are in place.
What I learned
IA isn’t just about where content lives. It’s about whether users can predict where it should live.
