Redesign of TU Delft Repository
Project overview
Navigating institutional repositories can often be a frustrating experience. Add to this the specific expectations of academic users, metadata standards, and legacy systems, and you get a complex ecosystem that demands thoughtful design. I joined the Research Infrastructure department of the TU Delft Library to improve the search experience of the repository and enhance the accessibility of the university's open research output.
My Role
As a UX/UI designer & researcher, I was responsible for the entire design process, including user research, conception, prototyping, visual design, and interaction development. I also worked as an “interim” Product Owner, overseeing the prioritization of new releases and administration of user stories.
Tools
Figma: For creating interactive prototypes and final designs.
Framer: For animating and visualizing user interactions.
Qualtrics: For generating quantitative user research insights through surveys.
Dovetail: For analyzing qualitative user research from usability testings.
Microsoft Clarity: For validating design and interactions.
Duration
November 2024 - Apr 2025
UX Audit
To gain a solid understanding of the task at hand, I conducted a UX audit of the existing Repository website. The site had an appealing design, which is important since good design that is user-friendly creates trust with users.
However, the UX of the website was complex and did not align with users’ heuristics, especially the search experience.
The search experience made it hard for users to find the most “relevant” document expected from the search query across all records in the repository. This was detrimental, since one of the most important features of a repository is to be able to find information as easily as possible.
Preliminary User Research
After the UX audit, I dived into the analytics of the website. Unsurprisingly, I found that since the launch of the latest website version (July 2024), user engagement and scrolling events had decreased considerably.
The analytics also revealed that the search experience negatively affected the user flow, as users often returned to the Home page when they couldn’t find what they were looking for.
Additionally, an initial CSAT survey showed a low satisfaction score of just 21%, clearly indicating that users were not happy with the search experience at all.
“The search function is bad, it is extremely hard to find articles and PhD thesis that previously were very easy to find” - Survey open response
Next up, I organized usability tests and workshops with a few users in order to measure the current search experience qualitatively. The goal for this session was to identify pain points, confusion, and opportunities for the redesign of the search experience.
The tests showed that the search experience was completely misaligned with what the users expected, which signaled the main challenge of the project.
Problem definition & UX KPIs
Based on the aforementioned research methods, the project faced the challenge of redesigning the search experience to meet users expectations. The central challenges included:
Increase user satisfaction: Ensuring that the CSAT score increased from 21% to at least more than 60%
Showing expected results: Improve search experience through the search bar by showing the expected results on at least the second row.
Reduce search time & cognitive load: Reduce time spent searching and try to implement search automation to reduce cognitive load.
The goal was to create a solution that not only meets the needs of the direct users but also fundamentally improves the convenience, efficiency, and accessibility of the TU Delft research.
Concept development
Building on the insights from the user research stage, I applied user-centered design principles to the redesign of the institutional repository. Guided by findings from user research and usability testing, I focused on resolving specific issues that hindered search efficiency, such as unclear filtering options, overwhelming result presentations, and inconsistent metadata.
Design process
In the design process, the focus was on creating a visually appealing and user-friendly interface that offers users an intuitive experience. Various design elements were carefully selected to convey a clear brand identity while maintaining a focus on accessibility, usability, and functionality.
I implemented a cleaner and more responsive interface, introduced more intuitive filtering and sorting mechanisms, and enhanced the consistency and visibility of metadata. These improvements streamlined user interactions and made it easier for users to locate and access relevant resources, resulting in a more seamless and satisfying search experience.
Final Impact and Learnings
Over a period of 5 months, with a new search improvement launching every few weeks, we achieved a CSAT increase from 21% to 70%.
As part of the search improvement strategy, I designed a new search bar and homepage that respected the initial design aesthetics. The goal was to modernize the visual aesthetic while keeping the vibrant TU Delft colors and brand.
The resulting design solution featured a tab system and a pre-filter dropdown that included all record types, such as master theses and student reports.
Final Impact and Learnings
Over a period of 5 months, with a new search improvement launching every few weeks, we achieved a CSAT increase from 21% to 70%.
As part of the search improvement strategy, I designed a new search bar and homepage that respected the initial design aesthetics. The goal was to modernize the visual aesthetic while keeping the vibrant TU Delft colors and brand.
The resulting design solution featured a tab system and a pre-filter dropdown that included all record types, such as master theses and student reports.