A redesign of MIT Libraries' website
Case Study Summary:
As one of the world’s leading research institutions, MIT libraries connect students, faculty, staff, and researchers to a vast range of scholarly resources, research tools, and archival collections. The goal of this project was to better understand the user experience on the Libraries' website and reimagine concrete, user-centered designs of the homepage and information architecture.
My role
As the lead researcher, I led the development and analysis of our research plan. I also facilitated team meetings, successfully managed relations with our client, and redesigned the homepage.
Teammates
Shreya Lohakare
Myra Chen
Aswathi Thilak
Project Type
Project Type
Information Architecture
Website Redesign
Usability Test
Tools
Figma
Userlytics
Optimal Workshop
Google Sheets
Zoom
Timeline
February- May 2025 (12 weeks)
Defining the problem
Problem statement
MIT Libraries recognizes the variety in experience and confidence that their students, faculty, staff, and researchers have with navigating their website. Their navigation menu can feel overwhelming and the catalog search features confusing.
How might we
How might we develop user-centered design recommendations for a website and information architecture redesign to make the search discovery experience more understandable for all users?
Two disability models​ for understanding a person’s ability to function
Researcher, Marno Retief (2018) defines the two as:
Medical Model
Highlights the limitations of a person’s disability and aims to fix it. A common "treatment" is the cochlear earplant.
Social Model
Societal changes improve the conditions or environments that contribute to a person’s ability to function.
Applying the Social Model
Looking to technology and travel apps to level the playing field for DHH travelers.
The process
Understand, explore, and materialize
To undertake this project, we undertook a major research analysis of the existing website to better understand the current user experience and opportunities to explore for a redesign. A final homepage and information architecture redesign was developed.
Two disability models​ for understanding a person’s ability to function
Researcher, Marno Retief (2018) defines the two as:
Medical Model
Highlights the limitations of a person’s disability and aims to fix it. A common "treatment" is the cochlear earplant.
Social Model
Societal changes improve the conditions or environments that contribute to a person’s ability to function.
Applying the Social Model
Looking to technology and travel apps to level the playing field for DHH travelers.
The problem
Lost in the stacks
The library website’s homepage and navigation system were described as overwhelming, cluttered, and unintuitive. Important resources, like research guides, citation tools, and library hours, were hard to locate.
Additionally, scholarly information seekers highly rely on search bars to find research topics, known articles, or general website information such as library hours. However, our research found that the currently existing search bar caused confusion, especially around the differences between Bento (pulls from multiple databases) and Primo search (pulls from advanced search), and given the lack of visibility around a site search function.
For context, Bento searches allow users of the website to search from a larger catalog of the library’s services into a “bento-like” search results format. Primo searches pull from a narrower library catalog that, unlike Bento, doesn’t include resources such as Archives and Theses. As a result, users frequently defaulted to the chat function or turned to Google Scholar to find answers, bypassing the site altogether.
Understanding the problem
Understanding the problem: conducting research & synthesizing insights
To fully understand how users interact with library websites, we conducted the following:
Literature review of university and public library websites highlight three user themes
15+ academic sources helped us frame problems around library literacy, information-seeking behaviors, and UI clarity.
A review of the literature on library websites surfaced three main topics for us to consider in our research: a pattern of information overload, users lacking library literacy, and users in search of greater scholarly support from institutions. These themes were found to overlap extensively with the generative research we conducted.
Competitive analysis highlight standards for clear labeling, easily discoverable services, and responsive designs
We evaluated 20 other library websites for best practices in navigation, accessibility, and search.
Our findings indicated trends across multiple university library websites that offer intuitive search experiences that differentiate between a site search and catalog search, navigation bars that are clearly labeled and therefore easy to use, offer easy-to-discover library services, and ideally, responsive designs across mobile and desktop experiences.
Stakeholder interview insights
Eight information seekers and four library liaisons shared frustrations, strategies, and workarounds. The research showed two main insights:
Issue 1
Divergence from familiar user search experiences
Everyday, web search behaviors and expectations negatively impact user experience in the Libraries’ website. Users expected the single search function on the existing library website to work like a Google search and were confused to find that to not be the case. Confusing nomenclature makes navigating a challenge.
Issue 2
Confusing jargon impedes site navigation
Many users found the labels and tools listed in the navigation bar of the website to be unfamiliar and therefore did not know how to fully navigate or utilize all the tools offered by the website.
Card sorting and tree testing highlighted task-based grouping, low confidence in navigation, and overwhelming library jargon by users
To re-invision the site’s information architecture, 20+ participants analyzed how users understood and grouped navigation terms.
Our findings indicate that users usually group navigation labels based on familiar tasks and goals, but they struggled to understand labels that used library or research-specific terms or lacked enough context.
Additionally, while most participants eventually located the correct information and completed the given tasks in our tree tests, low directness and confidence scores indicated that the navigation experience was not intuitive but involved lots of guesswork.
Two disability models​ for understanding a person’s ability to function
Researcher, Marno Retief (2018) defines the two as:
Medical Model
Highlights the limitations of a person’s disability and aims to fix it. A common "treatment" is the cochlear earplant.
Social Model
Societal changes improve the conditions or environments that contribute to a person’s ability to function.
Applying the Social Model
Looking to technology and travel apps to level the playing field for DHH travelers.
Usability testing showcased how prevailing mental models hinder discovery
We ran eight moderated and unmoderated tests on the current university libraries website and our redesigns.
Here, we found further evidence of users using the existing library website’s catalog search function as a site search. Additionally, users quickly and regularly relied on features allowing them to connect directly with a staff member, suggesting a lack of confidence and understanding of how to navigate the current website and its information architecture.
Key research insights
In agreement with academic research findings, information overload, library literacy and clearly labeled search functions were found to be important to consider. As a result, these insights shape our design priorities and process.
1
Terminology was overly academic or vague, confusing many users.
2
The Bento search bar (which pulls from multiple library catalogs, including archives and published theses) was often mistaken for a full-site search.
3
Overwhelming navigation options led to hesitation and task failure.
4
The search bar and staff support became fallbacks when users felt lost.
Guiding strategies for a redesign
-
Provide separate, clearly labeled site search and catalog search on the homepage.
-
Condense navigation categories to reduce overload and improve scanability.
-
Use clear, task-oriented language and improve labeling.
-
Make research support, hours, and help tools easier to find.
-
Add a floating live chat for immediate assistance.
Materialize
New information architecture
We also developed two IA models—a task-based and a topical version—and tested both to see which best aligned with user expectations.
Our two IA approaches aim to represent two potentially user mental models that may be interacting with a library website: one that is categorical and the other action-oriented.
Topical IA
Given that multiple competition library websites used clear and predictable navigation based on different topics, we redesigned a second IA approach to be “topical.” Topic-based navigation categories included search, services, and research.
Task-based IA
Our card sorting discovered that users organized labels into categories based on tasks or goals. We developed and named our first IA approach “task-based.” Tasks in this approach included starting research, finding library services, visiting the libraries, and news and events.
Two disability models​ for understanding a person’s ability to function
Researcher, Marno Retief (2018) defines the two as:
Medical Model
Highlights the limitations of a person’s disability and aims to fix it. A common "treatment" is the cochlear earplant.
Social Model
Societal changes improve the conditions or environments that contribute to a person’s ability to function.
Applying the Social Model
Looking to technology and travel apps to level the playing field for DHH travelers.
Another round of tree testing suggests the alignment of a Topica IA with user expectations.
Our findings suggest a topical IA approach more aligned with user mental models while users seemed unfamiliar with using a navigation bar based on tasks.
Topical IA
78%
task success
Topical vs original IA
7%
increase
Materialize
Redesigning the homepage
A variety of homepage designs were presented to the client to offer a range of possibilities for them to consider. Different components from each design version could then be picked and combined in the client’s future website redesign process and relevant stakeholder conversations. Showcased here is the version I designed based on our research findings.
My design: clear dual search, research guide visibility, and options for getting in touch.
Search function offers two tabs to search by: known items or exploratory
Spotlighting research guides and events
Featuring library services
Maintaining the existing position of library hours & locations
The impact
Navigation success:
+10%
​Search tool confidence:
+24%
In conclusion
Next steps
Designing within boundaries
With just 12 weeks, we had to prioritize features with the highest usability impact. Related to this, we a limitation in time and recruitment infrastructure limited our ability to test direct MIT Libraries' users.
Client feedback highlights wins and further opportunities
Although this project leverages public anecdotes from disabled individuals for redesign proposals, researchers Kelly Mack and Sophie Tian recommend that professionals must fully immerse themselves in Deaf culture, through ASL or Deaf Studies literature, to genuinely build empathy for the D/deaf and Hard of Hearing community (2020).
In particular, they liked the different ways of thinking about searching (site search vs how we label different tabs for the catalog search function), and the different ways users can find services and research experts.
Other comments we took into our final redesign included rethinking some word choices (such as “Library Services” or “liaisons,” as not everyone may be familiar with such terms). Links to our final IA and designs can be found at the bottom of the page.
Where it goes from here
Further validation of our redesigns, especially user testing with MIT’s user population, would benefit the continual improvement on MIT Librarie’s website and the users they serve.
Additional data metrics to consider tracking to measure impact include homepage bounce rates, search feature usage, and support requests.
Lastly, as of April 2026, MIT Libraries has decided to prioritize a redesign and restructuring of their site's search function. As such, users can now utilize their "universal search," which pulls from all of their databases. This function fully aligns and pulls from the findings of this project.