Democratizing UIs Through Meaningful Personalization
Role
Lead Researcher
Duration
2020 – 2025
Read Time
12 min
How could we enable end-users to completely redesign the interfaces they use?
Mainstream UI personalization features are typically limited to superficial aesthetic changes (like color mode). This research aimed at reimagine UI personalization: how could we enable end-users (anyone) to personalize any aspect of the UIs they use?
I identified critical pain points (e.g., effort, ideation, and expertise), elaborated requirements (e.g., data agency, ideation support, and execution assistance), and explored different prototypes. This enabled me not only to propose a solution but also to create a roadmap, a design space that developers of future personalization mechanisms should follow to ensure their solutions align with users' needs.
I followed a user-centered approach, interacting with more than 150 participants across 4 studies.
Discovery & Definition
I began with an in-depth qualitative study involving seven participants. Through three weeks of diary entries and co-design sessions, I immersed myself in their daily frustrations to uncover where current UI personalization falls short. To ensure these insights scaled, I then conducted a quantitative survey on Prolific. I found a significant demand for efficiency-oriented features and two critical points to address: the control-effort trade-off and the privacy-convenience paradox.
Interviews
Diary Studies
Focus Groups
Thematic Analysis
Surveys
Quantitative Analysis
Empathy Mapping
Personas
Ideating & Prototyping the Solution
After an ideation process, we decided to explore a 'request and assist' community-based personalization model to bridge the gap between 'Passive Adapters' and 'Power Users.' Using a functional prototype, GitUI, I conducted a two-week evaluative study with nine participants. Users were highly engaged with the community (82% request fulfillment) but identified ideation (imagining what to personalize in the first place) as a key barrier to long-term adoption.
Use Cases
Prototyping
Usability Testing
User Logging
Think-Aloud
Thematic Analysis
Speculative Design & Interaction Data
To explore solutions for the ideation gap, I used speculative design and experimental vignettes to study how 12 participants would engage with their own interaction data (e.g., clickstreams) to drive personalization. I found that system-initiated visual suggestions (previews of changes) that highlight long-term time savings are the most desired solution for solving the ideation challenge.
Speculative Design
Prototyping
Design Probes
Vignette Studies
Interviews
Affinity Diagramming
Delivery
I produced an End-User Personalization Design Space, a 7-dimension framework that product teams can use to introduce user-centered personalization features into everyday software. I also delivered GitUI, a functional web personalization tool featuring interaction data visualization (heatmaps).
The "Immutable Interface" and Ideation Pain Points
Most users perceive UIs as "immutable entities". They adapt their own behavior (creating inefficient workarounds) rather than adapting the interface. Simultaneously, ideation challenges prevent users from envisioning alternative layouts.
The "Zombie Mode" Kryptonite
Users often operate in a "zombie mode" of repetitive routines. Providing access to interaction data (heatmaps/clickstreams) supports reflection, making habits visible and shifting users from passive consumers to proactive decision-makers.
The Paradox of Automation
Users resist fully automated solutions due to loss of control. However, they prefer assisted solutions over purely manual ones. The "sweet spot" is software that proposes changes (visual previews of alternative UIs that can be further personalized) with a clear rationale, leaving the final decision to the user.
4 HCI Researchers
1 Engineer
1 Designer
Publications
Four top-venue
Thesis
PhD awarded (2026)
Design Space
A 7-dimension framework
Artefacts
Browser extension (GitUI)