Yuri Minami
Faster, Effortless Research Resource Discovery hero image

Faster, Effortless Research Resource Discovery

Uncovering how digital-native students discover research resources on the library website and addressing their needs

Project Type
UX consulting for School of Visual Arts Library
Focus
UX research & recommendation
My Team & Role
4 UX Consultants including me
Impact
Delivered 3 actionable recommendations that the client confirmed were system-widely applicable.

SVA research guides are a starting point of students' research

The School of Visual Arts (SVA) is a leading art and design college in NYC. Its library offers 44 unique research guide pages to help students start their research with curated resources and tips across various subjects.

Examples of SVA Library research guide pages
Some guides present lists of resources or archives, while others are tailored to specific classes, majors or topics. Their layouts also vary.

Library struggles to make guides self-serve and widely used

SVA Library faced a high volume of inquiries and aimed to make its services more self-serve, reducing reliance on direct librarian support. It also hoped to increase effective use of research guides, but a lack of insight into how users navigate the site made it challenging to improve.

Uncover how users navigate and find resources on research guides to improve their journey

Restricted by school-wide platforms & guidelines

As part of the school-wide website, SVA Library had limited control over page components, functionality, and style. These constraints had to be considered to ensure our UX recommendations' feasibility.

Choose priorities of guides and use cases to maximize impacts

Each research guide covered different topics, with varying layouts and purposes. The users and use cases are also diverse from students to faculties. We had to consider how to make the most impact within our limited time.

Used behavioral research methods to understand how users navigate

Used behavioral research methods to understand how users navigate.

Eye-Tracking

We chose eye-tracking because it is well-suited to the client's need to get insights into natural user behavior.

Retrospective Think-Aloud (RTA)

After seeing natural behavior using eye-tracking, we conducted interviews, using the RTA methods to gain insight into behaviors observed.

Metrics

  • Task completion rate (for checking effectiveness)
  • Time on task (for checking efficiency)
  • System Usability Scale (for checking satisfaction)

Participant

We recruited college students, who are the primary users, through personal networks, school mailing lists and on-site recruiting.

Sessions

We conducted 12 sessions (8 desktops and 4 mobiles).

Tested 2 high-priority pages in realistic academic research scenarios

To deliver the most impact within a limited timeframe, we selected the most frequently used and easily templatized pages for our testing scope. We designed tasks to reflect early-stage research scenarios, aligning with the guide's intent. In both tasks, participants were asked to find resources relevant to a prompt using the descriptions on the pages.

Online resources research scenario

1. Online Resource Page

This page showcases the databases available to SVA students.

Why this page?

As the main entry point with the highest traffic, this page had the greatest potential for impact.

Research Question

How effectively can users find, understand, and use the resources for their research needs?

Scenario & Task

The task involved finding a fashion magazine resource for a paper on the history of American fashion.

Design archives research scenario

2. Design Archives Page

This page provides a list of design archive resources with descriptions.

Why this page?

We asked the client to identify a key page that could be templatized for broader use, and they selected this page.

Research Question

What's the optimal way to organize individual research guides for clarity and ease of navigation?

Scenario & Task

The task was to find a typography resource for a movie poster design as a graphic design student.

Triangulated multiple data sources to derive findings

As a first step in our analysis, we used a rainbow sheet to synthesize common behavioral patterns and RTA feedback, giving us a broad overview of the core findings. Next, we triangulated quantitative metrics, eye-tracking data, and user quotes to extract insights and craft the story. Eye-tracking revealed what users actually did; quotes explained why they behaved that way, uncovering their needs; and quantitative metrics quantified the impact and severity of each issue.

Rainbow sheet used to synthesize SVA testing findings
Rainbow sheet led us to the core findings.

Most users could find helpful resources, but the experience was frustrating & inefficient

While 79% of students could complete the tasks, SUS score of just 41.3 indicates the user satisfaction of the research guide pages is poor. We identified 2 major UX issues each for the online resources and design archives pages.

79%

Task completion rate

41.3%

System Usability Score

Discovering relevant databases is difficult without prior knowledge due to search and skim difficulties

Discovering relevant databases is difficult without prior knowledge due to search and skim difficulties.

Insight

Students prefer to use the search bar to find the relevant databases

In the testing sessions, we identified that users prefer to use the search bar because they expected the search bar could show them the related resources sooner, like they experience in other websites.

67

%

Mentioned the preference for search

50

%

Searched immediately during the task

2.37

min

Average time on task

I would wish that the search bar can just immediately take me there. I don't want to spend time looking at everything.

Heatmap of online resource page
Heatmap of online resource page (Absolute duration).
Insight

Despite preferring the search bar, some students still couldn't get a relevant database result using it

Students often try to search, but the current functionality fails to return relevant databases when keywords are complex, such as those with multiple words or typographic errors.

67

%

Got NO matching database result from search bar (among participants who tried search)

Search returning no results for a complex keyword
When search keywords are complex, such as multiple words or containing typographical errors, no results are returned.
Insight

Students struggled with information overload when skimming

Instead of using the search bar, students attempted to skim the page, but the dense text and lack of information hierarchy made this difficult.

67

%

Mentioned information overload issue

I'm someone who's not very like patient, especially text. So this amount of, small text, I don't want to read it.

Eye-tracking result showing skipped paragraphs
Students tended to skip reading the paragraphs.

Make content easier to skim by improving the visual hierarchy and search bar experience.

First, I recommended fixing the search bar functionality. As a quicker, easier improvement, I also suggested making the content more skimmable to help students understand each database and identify those relevant to their research. Specifically, I proposed increasing logo size, enhancing the font size and weight of database labels, adding short keywords, and reducing descriptive text.

Online resource page before visual hierarchy recommendation
Before

The lack of visual hierarchy and excessive text made the content difficult to skim.

Online resource page after visual hierarchy recommendation
Recommendation

Visual hierarchy and reduced text help minimize information overload, allowing users to quickly skim database overviews using keywords.

Full recommendation design for making online resources easier to skim

Improve the search bar to support varied and complex queries.

I recommended enhancing the search functionality to be more robust, regardless of the keywords used, for example, by better handling complex queries and correcting typographic errors.

Search recommendation supporting multiple keywords
Recommendation

Allow multiple keywords in the search bar.

Search recommendation supporting typographic errors
Recommendation

Make search bar robust to typographic errors.

Resources in the third column and below the fold often go unnoticed

Resources in the third column and below the fold often go unnoticed.

Insight

The column labels are too small and vague to distinguish between categories

Many students navigated the page without grasping how columns were divided by category. The labels were too small to see, and the categories themselves were unclear and sometimes overlapped, offering no reliable cues to help users focus their search.

83%

Mentioned the issues related to the column labels

It isn't clear how these columns are separated, and sometimes their categories overlap.

Design archives page with 5 categories in 3 columns
The page is divided into 5 categories and structured using 3 columns.
Insight

Lack of category focus in 3-column layout caused inefficient exploration & attention gaps between columns

Because there was no clear category focus, many users scrolled inconsistently between columns. They often had to scroll back up multiple times to catch skipped resources, affecting efficiency. More importantly, the layout created unexpected attention gaps, with the third column receiving less attention than the first two columns.

It was like shopping at a market. I had to go from column to column.

Gaze replay from a participant showing inefficient unfocused column-to-column exploration.
Attention gaps in the design archives 3-column layout
The design archives page uses a 3-column layout intended to give each column equal importance, but it unintentionally drew uneven attention, making the third column less noticeable.

Make category groups noticeable and intuitive. Use a 1-column tab for linear focused browsing

We recommended using the existing tab component to group resources, making categories more visible and ensuring feasibility. We also advised switching to a single-column layout so users don't have to jump between columns, which enhances efficiency and eliminates unexpected attention gaps.

Recommendation design using tabs and a single-column layout

Successfully extended impact across guides, beyond scope

SVA project final presentation

At the end of the project, we presented our research findings and recommendations to the client and shared Slide deck, Problem list, Video reel, and Visual summary. They found our insights both interesting and helpful. Despite website constraints and a limited timeframe, we successfully focused on high-impact areas and delivered recommendations applicable to other research guides, as reflected in the following client quote.

We can use these suggestions to apply system-wide updates to all of our guides.

SVA Library Librarian

What I learned

Strengthened data storytelling

This project helped me connect quantitative metrics, eye-tracking behavior, and user quotes into a clear UX story. I want to keep using this skill to make research insights easier to act on as a product designer.

Next step: evaluate recommendation impact

I would like to test whether the recommendations improve the experience through follow-up evaluation methods such as A/B testing.