Enhancing Search for Seamless Discovery for 350M+ SonyLIV Users
Enhancing Search for Seamless Discovery for 350M+ SonyLIV Users
SonyLIV serves 350M+ users, and search is the gateway to discovering Originals, Live Sports, Movies, and TV Shows. But the legacy search experience created friction, people struggled to find relevant results, refine queries, and navigate a massive library.
SonyLIV serves 350M+ users, and search is the gateway to discovering Originals, Live Sports, Movies, and TV Shows. But the legacy search experience created friction, people struggled to find relevant results, refine queries, and navigate a massive library.
Redesigned SonyLIV Search to be faster, clearer, and more intelligent—scaling across Android, iOS, tablets, and low-end devices
Redesigned SonyLIV Search to be faster, clearer, and more intelligent—scaling across Android, iOS, tablets, and low-end devices
Role: Product Designer (End-to-end)
Role: Product Designer (End-to-end)
Role: Product Designer (End-to-end)
Platform: Fintech & AI, B2B SAAS
Platform: Fintech & AI, B2B SAAS
Platform: Fintech & AI, B2B SAAS
Team: 1 Designer, 1 PM, 1 PO, 3 Engineers
Team: 1 Designer, 1 PM, 1 PO, 3 Engineers
Team: 1 Designer, 1 PM, 1 PO, 3 Engineers
TL;DR
TL;DR
TL;DR
Problem
Problem
Legacy search made it hard to surface relevant results, refine queries, and navigate content at scale.
Legacy search made it hard to surface relevant results, refine queries, and navigate content at scale.
What I did
What I did
Redesigned the experience using failure patterns, user complaints, and global benchmarking—then validated changes through prototypes and quick feedback loops.
Redesigned the experience using failure patterns, user complaints, and global benchmarking—then validated changes through prototypes and quick feedback loops.
Impact
Impact
Improved search-to-play conversion, reduced time-to-result friction, increased search success rate, boosted retention during sports events, and improved performance for low-end devices.
Improved search-to-play conversion, reduced time-to-result friction, increased search success rate, boosted retention during sports events, and improved performance for low-end devices.
Context (Set the stakes)
Context (Set the stakes)
Context
Search is a high-intent moment. People use it when they already know what they want, or when they want the fastest path to something worth watching. When search fails, users don’t browse longer they leave.
Search is a high-intent moment. People use it when they already know what they want, or when they want the fastest path to something worth watching. When search fails, users don’t browse longer they leave.
The Problem
The Problem
The Problem
The legacy experience broke in predictable ways:
Relevance wasn’t consistent
Refinement was hard
Results were difficult to scan
Performance issues hurt low-end devices globally
The legacy experience broke in predictable ways:
Relevance wasn’t consistent
Refinement was hard
Results were difficult to scan
Performance issues hurt low-end devices globally
Before: search made users work too hard to get to a playable result.”
Before: search made users work too hard to get to a playable result.”
Goals & Success Metrics
Goals & Success Metrics
Goals & Success Metrics
Goals
Goals
Make search faster and clearer
Improve relevance and discoverability.
Support refinement (filters/categories) without overwhelming users
Make the experience reliable across devices and network quality
Make search faster and clearer
Improve relevance and discoverability.
Support refinement (filters/categories) without overwhelming users
Make the experience reliable across devices and network quality
Success metrics (what we aimed to move)
Success metrics (what we aimed to move)
Search-to-play conversion
Time-to-result friction
Search success rate (typed + suggested)
Retention during high-traffic sports events
Performance on low-end devices / low bandwidth
Search-to-play conversion
Time-to-result friction
Search success rate (typed + suggested)
Retention during high-traffic sports events
Performance on low-end devices / low bandwidth
What I owned + collaboration
What I owned + collaboration
What I owned + collaboration
I led end-to-end Search UX/UI including: UX principles, competitive research, IA & categorization, insight mapping, interaction/UI design, responsive adaptation, usability testing, and handoff specs.
I led end-to-end Search UX/UI including: UX principles, competitive research, IA & categorization, insight mapping, interaction/UI design, responsive adaptation, usability testing, and handoff specs.
What's been learned (Discovery → Insights)
What's been learned (Discovery → Insights)
What's been learned (Discovery → Insights)
My process started with why search was breaking, before designing how it should work:
My process started with why search was breaking, before designing how it should work:
Identified drop-offs and failure patterns
Extracted recurring complaints from reviews and user interviews
Benchmarked global patterns (Netflix, Amazon Prime, YouTube)
Clustered issues into accuracy, discoverability, relevance, refinement
Used these clusters to guide UI decisions
Prototyped and validated through quick feedback loops
Identified drop-offs and failure patterns
Extracted recurring complaints from reviews and user interviews
Benchmarked global patterns (Netflix, Amazon Prime, YouTube)
Clustered issues into accuracy, discoverability, relevance, refinement
Used these clusters to guide UI decisions
Prototyped and validated through quick feedback loops
The 4 insight themes
The 4 insight themes
Relevance breaks trust — if the first screen feels wrong, users give up.
Discoverability needs structure — users need helpful suggestions and clear categories.
Refinement must be effortless — filters and sorting should reduce effort, not add it.
Performance is part of UX — search must feel quick, especially during sports spikes and on low-end devices.
Relevance breaks trust — if the first screen feels wrong, users give up.
Discoverability needs structure — users need helpful suggestions and clear categories.
Refinement must be effortless — filters and sorting should reduce effort, not add it.
Performance is part of UX — search must feel quick, especially during sports spikes and on low-end devices.
Key Decisions
Key Decisions
Key Decisions
Decision 1: Make “first results” scannable and structured
Decision 1: Make “first results” scannable and structured
Why
Why
Users decide in seconds whether search is “working.”
Users decide in seconds whether search is “working.”
What changed
What changed
Clear result layout + metadata that supports fast selection.
Clear result layout + metadata that supports fast selection.
Decision 2: Reduce typing effort with dynamic suggestions
Decision 2: Reduce typing effort with dynamic suggestions
Why
Why
Faster input increases success on the first try.
Faster input increases success on the first try.
What changed
What changed
Smart suggestions that guide users while typing.
Smart suggestions that guide users while typing.
Decision 3: Make refinement obvious (filters + categories)
Decision 3: Make refinement obvious (filters + categories)
Why
Why
Libraries are massive—users need control without friction.
Libraries are massive—users need control without friction.
What changed
What changed
Cleaner categorisation + better filters to narrow quickly.
Cleaner categorisation + better filters to narrow quickly.
Decision 4: Optimize for low-end devices and low bandwidth
Decision 4: Optimize for low-end devices and low bandwidth
Why
Why
Performance directly affects search completion and retention.
Performance directly affects search completion and retention.
What changed
What changed
Responsive layouts + performance-first design decisions across devices.
Responsive layouts + performance-first design decisions across devices.
Solution Walkthrough
Solution Walkthrough
Solution Walkthrough
Impact
Impact
Impact
Improved search-to-play conversion (better relevance + scannability)
Reduced time-to-result friction (suggestions + clearer refinement)
Increased search success rate for typed + suggested queries
Boosted retention during sports events (real-time sports visibility)
Enhanced experience for low-end devices / low bandwidth (performance-first UI)
Improved search-to-play conversion (better relevance + scannability)
Reduced time-to-result friction (suggestions + clearer refinement)
Increased search success rate for typed + suggested queries
Boosted retention during sports events (real-time sports visibility)
Enhanced experience for low-end devices / low bandwidth (performance-first UI)
Outcome
Outcome
Outcome
Use your existing outcome sections as the closing story:
Use your existing outcome sections as the closing story:
Improved search efficiency (fewer refine loops, less typing)
Higher search-to-play conversions (structured layouts + ranking)
Increased user satisfaction (reviews + clarity + filters)
Stronger retention during peak events (handled spikes smoothly)
Optimised for all devices (responsive + reduced lag)
Improved search efficiency (fewer refine loops, less typing)
Higher search-to-play conversions (structured layouts + ranking)
Increased user satisfaction (reviews + clarity + filters)
Stronger retention during peak events (handled spikes smoothly)
Optimised for all devices (responsive + reduced lag)
Learnings + Next Steps
Learnings + Next Steps
Learnings + Next Steps
Learnings
Learnings
Search UX is about reducing uncertainty fast.
Performance isn’t a technical detail, it’s the product.
Search UX is about reducing uncertainty fast.
Performance isn’t a technical detail, it’s the product.
Next steps
Next steps
Personalised suggestions (based on viewing intent)
Better query understanding (spelling variants, multilingual)
Instrumentation by intent type (known-title vs browse-intent)
Personalised suggestions (based on viewing intent)
Better query understanding (spelling variants, multilingual)
Instrumentation by intent type (known-title vs browse-intent)