Search is high frequency.
Over half of participants reported using search multiple times per day, making even small inefficiencies meaningful over time.
Case Study / 02
Improving high-frequency search in Spotify’s mobile app.
Timeline
March 1 – April 8, 2026
Role
UX Researcher
Product Designer
Team
Individual HCI Project
Disciplines
Needfinding
HCI Evaluation
Prototyping
Mobile UX
Overview
Reducing friction in one of Spotify’s most frequent interactions.
This project investigated how Spotify’s mobile search experience could better support both known-item retrieval and exploratory discovery. Search is a primary interaction mechanism for locating songs, artists, playlists, podcasts, and other audio content, but the current flow can introduce small delays through tab switching, mixed result types, and repeated scanning.
I redesigned the experience around a persistent bottom search bar that keeps search accessible from the Home interface, reduces navigation overhead, and supports faster one-handed retrieval while preserving discovery.
Problem Overview
Spotify search is effective, but repeated micro-frictions add up for frequent mobile users.
Locating audio content often requires navigating to a separate Search tab, entering a query, and scanning across multiple result categories. These steps may seem minor, but they become meaningful in high-frequency and context-constrained situations such as commuting, exercising, studying, or multitasking.
How might we
reduce search-related interaction cost while preserving Spotify’s support for discovery and exploration?
Research Methods
Using mixed methods to understand Spotify search behavior.
I used a mixed-methods approach to understand where friction occurs in Spotify’s mobile search workflow and how search intent changes across contexts. Methods included a quantitative survey, semi-structured interviews, contextual apprenticeship with think-aloud observation, and heuristic evaluation of the existing interface.
View Survey Results.
01
Survey
Captured search frequency, perceived usability, frustration, speed, and context.
02
Interviews
Explored retrieval behavior, exploratory search, friction points, and workarounds.
03
Think aloud
Observed real-time search behavior, hesitation, scanning, and corrections.
04
Heuristics
Evaluated visibility, match with user intent, and efficiency of use.
Research findings
Search is fast for many users, but execution-level friction still appears.
The research revealed a gap between general satisfaction and repeated interaction costs. Users rated Spotify search positively overall, but interviews, observation, and heuristic evaluation showed that scanning, tab switching, and mixed content types still create friction during frequent use.
Intent changes the task.
Retrieval-dominant users prioritized exact matches and speed, while exploratory users needed stronger support for browsing, filtering, and comparison.
Mobile context matters.
Participants often searched while commuting, exercising, or multitasking, making one-handed access and reduced scanning especially important.
Needfinding insights
The opportunity was not a full redesign of Spotify, but a reduction of cumulative search friction.
Findings pointed to four design priorities: minimize micro-interaction costs, better distinguish retrieval from exploration, reduce redundant navigation between Home and Search, and support fast decision-making in constrained mobile contexts.
Ideation
Exploring structural changes to Spotify’s search flow.
Brainstorming focused on interaction architecture rather than visual styling. I generated twenty design ideas across four themes: navigation elimination, motor optimization, retrieval optimization, and exploration support. Three concepts were selected for low-fidelity prototyping.
01
Unified Home with Persistent Bottom Search
Integrates search into the Home interface with a bottom-anchored search bar, reducing the need to navigate to a separate tab.
02
Adaptive Unified Home Interface
Transitions dynamically between browsing and search states when the user begins typing, supporting retrieval and discovery in one flow.
03
Hybrid Home with Predictive Retrieval
Surfaces recent searches, frequently played content, and predicted recommendations to reduce reliance on manual search and support frequent actions.
Low-fidelity prototypes
Testing three navigation models before refining the final direction.
Each low-fidelity prototype represented a different way to reduce search friction: anchoring search at the bottom, dynamically transitioning between Home and Search, or using predictive retrieval shortcuts.
First evaluation
The persistent bottom search direction performed strongest.
Five participants evaluated the three low-fidelity prototypes through representative search tasks: finding a specific song, searching for a playlist by genre, and searching for an artist. Prototype 1 received the highest overall mean score and was perceived as the fastest and most intuitive option.
01
Prototype 1
Highest overall mean score at 4.60 and preferred by four out of five users.
02
Prototype 2
Balanced search and browsing but introduced ambiguity in state changes.
03
Prototype 3
Supported discovery but felt slower for known-item retrieval.
04
Direction
Refine persistent search visibility, bottom placement, and unified browsing.
Final solution
A persistent bottom search experience built into Spotify Home.
The final medium-fidelity prototype integrates search directly into the Home interface. Search remains fixed at the bottom of the screen, expands inline when activated, and supports both known-item retrieval and exploratory browsing without requiring users to switch to a separate Search tab.
01
Persistent bottom search
The search bar remains accessible at the bottom of the screen, improving reachability and reducing the repeated step of navigating to a separate Search tab.
02
Unified browsing and retrieval
Search expands within the Home experience, allowing users to move between discovery and goal-oriented retrieval without cognitive switching.
03
Prioritized result hierarchy
Search results prioritize exact matches before broader category options, reducing scanning effort during high-intent retrieval tasks.
Final prototype
A medium-fidelity prototype demonstrating the complete search flow.
The final prototype was created in Figma and demonstrated three representative tasks: finding a specific song, finding a playlist by genre, and finding a specific artist. The interaction highlights how persistent search access reduces navigation while maintaining browsing support.
Final evaluation results
Participants completed tasks quickly with minimal errors.
In the final evaluation, five participants completed representative Spotify search tasks using the medium-fidelity prototype. Total task completion time ranged from approximately 14.4 to 18.5 seconds, with most participants completing all tasks in 9 to 10 clicks. The average error rate was approximately 6%.
Qualitative feedback indicated that participants quickly identified the persistent search bar, found the layout intuitive, and felt the design reduced navigation effort. Minor confusion around predictive suggestions suggested a future opportunity to improve labeling and visual hierarchy.
14.4–18.5s
total task completion range
~6%
average error rate
Reflection