← All Work 06 / 06
Usability Testing Remote Research Digital Archives

Barnard Digital Collections Usability Study

Ten remote usability sessions on Barnard College's digital archive — identifying discoverability barriers and navigation failures, then delivering four prioritized recommendations with Figma mockups for immediate implementation.

Role
UX Researcher · Moderator · Designer
Team
Hiral Parekh, Lillian Gooden, Mohammed Hosen
Duration
7 weeks · Spring 2020
Tools
Adobe Illustrator · Photoshop · Zoom
Barnard Digital Collections
01 — Context

Testing a Long-Unchanged Archive

Barnard Digital Collections provides students and researchers with free access to nearly 85,000 academic sources — newspapers, scrapbooks, photographs, and archival documents. Despite the scale of the collection, no major design changes had been made to the website since its inception.

The client wanted us to test the desktop version of the site, with particular focus on two specific features: the filter/metadata system and the exhibits section. A website update was already in the pipeline — our study would inform exactly what to prioritize.

85k
Academic Sources
10
Remote Sessions
4
Final Recommendations
85%
Desktop Users

Client Goals

  • Enhance discoverability of the collection for better user engagement
  • Understand how users interact with the metadata (filter) categories
  • Assess whether the exhibits section is being found and used effectively
02 — Research

Remote Usability Testing via Zoom

Originally planned as in-person sessions, COVID-19 forced a complete pivot to remote testing via Zoom. Participants were asked to share their screen and think aloud throughout. Sessions were recorded for cross-referencing.

From a sample of 20 candidates, we shortlisted 8 — a deliberate mix of Barnard students, staff, and external researchers, some of whom had never used digital archives for research before. Each team member moderated 2 sessions; I moderated 2 and served as note-taker for another 2.

Scenario & Tasks

We presented all participants with the same scenario to ensure task comparability:

"Imagine that you are an independent researcher looking for information on student life at Barnard College in the 1960s and '70s. You have come to the Barnard Digital Collections portal to source materials from their archive."

— Study scenario given to all participants
Task 01
Find the digital exhibit on Student Publishing.
Task 02
Locate an article about Martha Stewart (class of 1964) in the Barnard Alumnae magazine.
Task 03
Search for proposals on second-wave feminists' concerns for women in higher education.
Task 04
Find photos of Greek games held at Barnard in 1966 using the digital collection filters.
Post Test Questionnaire
Post-test questionnaire responses — capturing overall experience and additional recommendations
03 — Analysis

High-Level Findings

After collecting data from all 10 sessions, each team member independently analyzed their interviews and compiled findings. We then converged to identify the patterns that cut across participants and select 4 final recommendations based on severity and implementation feasibility.

Finding 01
Overall aesthetic was well-received — users appreciated the colour scheme. But aesthetic satisfaction masked functional failures that only emerged during task completion.
Finding 02
Metadata filter categories were genuinely useful for narrowing results — but the sheer number of categories was overwhelming for most users.
Finding 03
Inexperienced users were unaware of advanced search operators ("AND", "OR", "NOT") — and the site provided no guidance on where to learn them.
Finding 04
"View Object" button in the exhibits section did not appear clickable — users skipped over actionable elements because they didn't register as interactive.
Figure 1 - Participant Types Figure 2 - Archive Experience
Figure 1: Participant breakdown by type | Figure 2: Prior digital archive experience
04 — Recommendations

Four Prioritized Design Solutions

Recommendation 1 — Added Contrast & Better Error Signifiers

Finding: 37.5% of users completely missed the "Narrow Your Search" panel. 75% of users struggled with the date filter — entering only a start date produced a vague error with no guidance on how to fix it.

Solution: Increase visual contrast on the "Narrow Your Search" panel to draw attention. Improve error messaging to clearly explain what's required, and add an information button to prevent the error altogether.

Recommendation 1
Mockup: Improved filter panel contrast and contextual error messaging

Recommendation 2 — Surface Advanced Search Operators

Finding: Inexperienced researchers were unaware that "AND", "OR", double quotes, and other search operators were available — the functionality existed but was completely hidden.

Solution: Add operator shortcut buttons directly above the global search bar, making them visible and accessible without requiring users to seek documentation.

Recommendation 2
Mockup: Advanced search operators surfaced at point-of-use

Recommendation 3 — Sub-navigation for Exhibits

Finding: The exhibits section only offered "Next" and "Prev" navigation — users had no sense of how many exhibits existed or where they were in the collection.

Solution: Add a secondary navigation menu listing individual exhibit titles under main section headings — giving users a map of the collection and a sense of progress.

Recommendation 3
Mockup: Sub-navigation providing context and progress within the exhibits section

Recommendation 4 — Show Photos in Full on First Click

Finding: Clicking a photo loaded a default cropped view — users thought they had navigated to a different photo entirely. Zoom controls were also poorly visible in the cropped state.

Solution: Display the full photo when first clicked, letting users choose to zoom in rather than defaulting to a crop. Restyle zoom controls to be visually prominent.

Recommendation 4
Mockup: Full image display on first click with prominent zoom affordance
05 — Outcome

Client Response & Learnings

The client was genuinely appreciative — particularly because our recommendations were easy-to-implement, progressive improvements rather than a full redesign. They had been aware of some of these issues but hadn't framed them as significant UX problems. The study crystallized the issues and provided implementable evidence.

The clients were glad to see that our recommendations were practical enough to be implemented through minor development changes — none required a structural overhaul of the site.

— Project outcome

Post-Presentation Exploration

After the client presentation, we explored the possibility of a horizontal filter layout for better discoverability. After producing a mockup, we concluded it introduced new problems — it was cluttered and required careful mouse navigation. A vertical dropdown with good labelling remained the better solution.

Horizontal Filter Exploration
Horizontal filter exploration — ultimately rejected in favour of the improved vertical approach
Learning 01
Remote testing via Zoom, when participants feel comfortable, can actually produce richer think-aloud data than in-person sessions where physical presence adds social pressure.
Learning 02
Post-meeting explorations are valuable — the horizontal filter concept would not have been ruled out without a concrete mockup to evaluate against the original design.
Learning 03
Clients already know about most issues — what research delivers is prioritization and evidence strong enough to justify the development investment of fixing them.
View Final Report →