CLIENT
Personal Concept
ROLE
UX Designer
YEAR
2024
STATUS
Completed
Overview
Overview
Overview
I lead a UX initiative to rescue a once-pioneering streaming platform (codenamed “Roku Reimagined”) from what seems to be an imminent slide relative to a key competitor. While Roku carved out a strong foothold in the external media device segment, consumer habits by 2024 increasingly favor smart TVs with advanced built-in OSs—like Google TV. Through in-depth user research and a double-pronged competitor benchmark, I set out to clarify precisely why Roku was losing ground and how a content-driven, AI-centric solution could address both user demands and competitor shortcomings.
I lead a UX initiative to rescue a once-pioneering streaming platform (codenamed “Roku Reimagined”) from what seems to be an imminent slide relative to a key competitor. While Roku carved out a strong foothold in the external media device segment, consumer habits by 2024 increasingly favor smart TVs with advanced built-in OSs—like Google TV. Through in-depth user research and a double-pronged competitor benchmark, I set out to clarify precisely why Roku was losing ground and how a content-driven, AI-centric solution could address both user demands and competitor shortcomings.
Defining the Problem
Defining the Problem
Defining the Problem
Roku was once synonymous with streaming—a “streamer in a box” that changed how people consumed media at home. By 2024, however, three critical challenges were found during the discovery phase:
Roku was once synonymous with streaming—a “streamer in a box” that changed how people consumed media at home. By 2024, however, three critical challenges were found during the discovery phase:
01
01
01
Shift from External Boxes to Built-In OSs
Shift from External Boxes to Built-In OSs
69%
69%
of consumers, according to surveys, found external boxes increasingly redundant, relying instead on integrated TV operating systems.
of consumers, according to surveys, found external boxes increasingly redundant, relying instead on integrated TV operating systems.
02
02
02
Competition with Google TV
Google TV soared to an estimated 270+ million active users (versus Roku’s 83 million), thanks to robust AI-driven curation and manufacturer integration.
While Roku itself continued to grow, Google TV’s rapid adoption—especially in built-in TV form—threatens to cannibalize the market. The reasons for such adoption weren’t always crystal clear, but the reasons became clearer with some exploration
Competition with Google TV
Google TV soared to an estimated 270+ million active users (versus Roku’s 83 million), thanks to robust AI-driven curation and manufacturer integration.
While Roku itself continued to grow, Google TV’s rapid adoption—especially in built-in TV form—threatens to cannibalize the market. The reasons for such adoption weren’t always crystal clear, but the reasons became clearer with some exploration
03
03
03
Stagnant UX
My interviews revealed growing frustration with Roku’s static grid layout, shallow “thumbs up/down” recommendations, and minimal cross-device synchronization—leading watchers to see Roku as outdated for a world demanding slick, content-centric solutions.
Stagnant UX
My interviews revealed growing frustration with Roku’s static grid layout, shallow “thumbs up/down” recommendations, and minimal cross-device synchronization—leading watchers to see Roku as outdated for a world demanding slick, content-centric solutions.
Core Objective
Core Objective
Create a next-generation UX that:
Reinvents Roku’s interface for modern tastes
Delivers advanced personalization, bridging the gap between user feedback and content discover
Encourages frictionless usage across devices and user scenarios
Exploits competitor shortcomings to build a truly differentiated experience
Create a next-generation UX that:
Reinvents Roku’s interface for modern tastes
Delivers advanced personalization, bridging the gap between user feedback and content discover
Encourages frictionless usage across devices and user scenarios
Exploits competitor shortcomings to build a truly differentiated experience
Create a next-generation UX that:
Reinvents Roku’s interface for modern tastes
Delivers advanced personalization, bridging the gap between user feedback and content discover
Encourages frictionless usage across devices and user scenarios
Exploits competitor shortcomings to build a truly differentiated experience
Deep-Dive Research
Deep-Dive Research
Deep-Dive Research
I aimed for a strategic, non-arbitrary approach to unearth user pain points and shape well-informed design decisions. Each step in my research yielded discoveries that directly influenced the proposed solution, while a double-pronged perspective let me weigh user frustrations alongside competitor strengths and weaknesses.
I aimed for a strategic, non-arbitrary approach to unearth user pain points and shape well-informed design decisions. Each step in my research yielded discoveries that directly influenced the proposed solution, while a double-pronged perspective let me weigh user frustrations alongside competitor strengths and weaknesses.
More specifically this double-pronged approach helped me:
Pinpoint user frustrations and desires in streaming (primary research: surveys, interviews, user journeys)
Benchmark Roku against competitors like Google TV (secondary research: adoption rates, brand synergy, manufacturer perspectives) to see where Roku could do better—and potentially exploit competitor weaknesses for a unique competitive edge.
More specifically this double-pronged approach helped me:
Pinpoint user frustrations and desires in streaming (primary research: surveys, interviews, user journeys)
Benchmark Roku against competitors like Google TV (secondary research: adoption rates, brand synergy, manufacturer perspectives) to see where Roku could do better—and potentially exploit competitor weaknesses for a unique competitive edge.
01
Secondary & Competitor Analysis
Kicking off with a comparative study, I realized Google TV had cornered the market on built-in OS integration, advanced AI, and brand synergy. Roku, meanwhile, continued to rely heavily on standalone devices and a static UI.
Secondary & Competitor Analysis
Kicking off with a comparative study, I realized Google TV had cornered the market on built-in OS integration, advanced AI, and brand synergy. Roku, meanwhile, continued to rely heavily on standalone devices and a static UI.
Brand synergy and flexible OS customization gave Google TV an edge with TV makers, while users praised its aggregated home screen and powerful recommendations.
Brand synergy and flexible OS customization gave Google TV an edge with TV makers, while users praised its aggregated home screen and powerful recommendations.
Although Roku’s own share of streaming devices remained strong (43% in the standalone segment), the overall market shift to more dynamic built-in streaming systems has left them vulnerable.
Although Roku’s own share of streaming devices remained strong (43% in the standalone segment), the overall market shift to more dynamic built-in streaming systems has left them vulnerable.
02
Survey Insights
Survey responses, primarily from 18–24-year-olds (79% of participants), revealed two crucial points:
Survey Insights
Survey responses, primarily from 18–24 year-olds (79% of participants), revealed two crucial points:
Survey Insights
Survey responses, primarily from 18–24-year-olds (79% of participants), revealed two crucial points:
Only Moderately Satisfying Recommendations
Only Moderately Satisfying Recommendations
Built-In TV Operating Systems are Popular
Built-In TV Operating Systems are Popular
This hammered home the need for rich, time-saving recs that cater to an audience eager for immediate, relevant content without hopping between apps.
This hammered home the need for rich, time-saving recs that cater to an audience eager for immediate, relevant content without hopping between apps.
03
Interview Synthesis
I conducted in-depth interviews spanning multiple age groups and usage patterns:
Interview Synthesis
I conducted in-depth interviews spanning multiple age groups and usage patterns:
External Rating Dependence
External Rating Dependence
Some users rely on Rotten Tomatoes or IMDB before committing to a show.
Some users rely on Rotten Tomatoes or IMDB before committing to a show.
Desire for a “Book Club” Feel
Desire for a “Book Club” Feel
Some watchers wanted a socially aware environment—where critiques and discussions thrive—to help them choose better content more easily.
Some watchers wanted a socially aware environment—where critiques and discussions thrive—to help them choose better content more easily.
Mixed Responses to Watch Parties
Mixed Responses to Watch Parties
While some loved the shared experience, others found it extraneous post-pandemic.
While some loved the shared experience, others found it extraneous post-pandemic.
04
04
04
Persona Snapshots
Personas emerged from these interviews and user data, each representing a key user type to help guide this process.
Survey Insights
Survey responses, primarily from 18–24 year-olds (79% of participants), revealed two crucial points:
Persona Snapshots
Personas emerged from these interviews and user data, each representing a key user type to help guide this process.
Alex Rivera
Alex Rivera, a 28-year-old software engineer in Seattle, enjoys binge-watching sci-fi, fantasy, and indie films. Despite being tech-savvy, Alex struggles with managing multiple subscriptions and overwhelming content. Offline viewing is crucial due to a busy schedule. Alex values intuitive interfaces but is frustrated by poor recommendations.
"I want a streaming service that knows me with perfect recommendations. Why is that so hard to find?"
Mark Kia
Mark Kia is a lifetime cable television watcher. He enjoys the convince of not having to think, and just turning the channel to his trusted networks. Since the cost of cable got high, his son set up the TV with Roku, He doesn’t like that he has to hunt and browse for things, it feels like a chore to him. He mostly likes watching re-runs of his favorite shows and finds it challenging to get what he wants right away.
"I just want to watch my shows without dealing with streaming apps. Cable TV has everything I need."
Identifying the Gap
Pulling from both prongs—user demands (aggregated, AI-driven recs) and competitor data (Google TV’s content-curation approach, manufacturer preference)—I identified the focal shortcoming.
Roku’s app-grid UI, lacking advanced curation, left them looking outdated. Simultaneously, Google TV didn’t always provide the deeper user feedback loop or optional community features some participants craved. This revealed an opportunity to design not just a “catch-up” solution but one that surpassed it
User Journey Mapping
I sketched out potential flows for advanced personalization:
User Journey Mapping
I sketched out potential flows for advanced personalization:
Diving Deeper Than “Thumbs Up/Down”: Users wanted to clarify why they like or dislike content, ensuring more accurate recs.
Conversational AI: Scenes where a user “talks” to the service about story preferences or disliked tropes, drastically refining suggestions over time.
Diving Deeper Than “Thumbs Up/Down”: Users wanted to clarify why they like or dislike content, ensuring more accurate recs.
Conversational AI: Scenes where a user “talks” to the service about story preferences or disliked tropes, drastically refining suggestions over time.
Ideation & Concept Development
Ideation & Concept Development
Ideation & Concept Development
Everything pointed to the same conclusion: Roku’s UI needed to transform from a dull app grid into a dynamic ecosystem that fosters user engagement and trust in its recommendations, bridging what Google TV succeeded at (fast, aggregated content) while addressing the desire for deeper user feedback and optional community features.
Everything pointed to the same conclusion: Roku’s UI needed to transform from a dull app grid into a dynamic ecosystem that fosters user engagement and trust in its recommendations, bridging what Google TV succeeded at (fast, aggregated content) while addressing the desire for deeper user feedback and optional community features.
Content-Centric Home Screen
Content-Centric Home Screen
Content-Centric Home Screen
Gone is the flat grid. Instead, feature aggregated lists from multiple streaming services, curated by user preferences, ratings, or trending content—directly on the home screen.
Gone is the flat grid. Instead, feature aggregated lists from multiple streaming services, curated by user preferences, ratings, or trending content—directly on the home screen.
Aggregated Ratings & Social Layer
Content-Centric Home Screen
Aggregated Ratings & Social Layer
Integrate external rating scores (IMDB, Rotten Tomatoes) while allowing an opt-in communal discussion forum. For those uninterested in social features, the UI remains minimal, preventing clutter.
Integrate external rating scores (IMDB, Rotten Tomatoes) while allowing an opt-in communal discussion forum. For those uninterested in social features, the UI remains minimal, preventing clutter.
Conversational AI
Provide an optional assistant that can parse user statements—e.g., “I like historical dramas but prefer comedic undertones”—to deliver more confident matches. This AI also evolves after each user feedback loop, bridging the subtle preferences that “thumbs up/down” can’t capture.
Provide an optional assistant that can parse user statements—e.g., “I like historical dramas but prefer comedic undertones”—to deliver more confident matches. This AI also evolves after each user feedback loop, bridging the subtle preferences that “thumbs up/down” can’t capture.
To minimize friction, I envisioned the AI operating in two modes:
To minimize friction, I envisioned the AI operating in two modes:
01 Active State
01 Active State
Users can deliberately activate the AI (voice or text) whenever they want to adjust preferences, clarify likes/dislikes, or ask for custom recs. This ensures the AI is intentionally engaged and doesn’t overwhelm those who prefer minimal assistance.
Users can deliberately activate the AI (voice or text) whenever they want to adjust preferences, clarify likes/dislikes, or ask for custom recs. This ensures the AI is intentionally engaged and doesn’t overwhelm those who prefer minimal assistance.
02 Passive State
02 Passive State
When watchers browse content and exhibit patterns (e.g., quickly exiting multiple shows or skipping similar genres), the AI can quietly analyze these behaviors. It may rearrange the home screen in real time, surfacing more relevant options based on time of day, day of the week, or any trending preferences observed.
When watchers browse content and exhibit patterns (e.g., quickly exiting multiple shows or skipping similar genres), the AI can quietly analyze these behaviors. It may rearrange the home screen in real time, surfacing more relevant options based on time of day, day of the week, or any trending preferences observed.
If it detects repeated dissatisfaction—like someone sampling two or three titles and bailing almost instantly—it can unobtrusively offer a brief conversation: “Not liking these action flicks tonight? Want help finding something cozier?” The user can choose to engage or dismiss it.
If it detects repeated dissatisfaction—like someone sampling two or three titles and bailing almost instantly—it can unobtrusively offer a brief conversation: “Not liking these action flicks tonight? Want help finding something cozier?” The user can choose to engage or dismiss it.
Minimizing Intrusion
Minimizing Intrusion
My design ensures the AI never blocks the primary UI content in passive mode. If it’s triggered passively, it appears as a discreet prompt near the screen edge, never overshadowing the show thumbnails or navigation menus. Moreover, the AI always confirms whether it’s being too intrusive and provides an easy toggle to deactivate passive suggestions altogether. This respects watchers like Mark, who might find constant guidance distracting, and embraces Alex’s desire for in-depth personalization.
My design ensures the AI never blocks the primary UI content in passive mode. If it’s triggered passively, it appears as a discreet prompt near the screen edge, never overshadowing the show thumbnails or navigation menus. Moreover, the AI always confirms whether it’s being too intrusive and provides an easy toggle to deactivate passive suggestions altogether. This respects watchers like Mark, who might find constant guidance distracting, and embraces Alex’s desire for in-depth personalization.
Rationale
Rationale
Active for those who want direct control, proactively telling the AI what they do or don’t like,
Seamless for watchers who appreciate subtle, on-the-fly adjustments—without having to manually switch anything on.
Active for those who want direct control, proactively telling the AI what they do or don’t like,
Seamless for watchers who appreciate subtle, on-the-fly adjustments—without having to manually switch anything on.
Active for those who want direct control, proactively telling the AI what they do or don’t like,
Seamless for watchers who appreciate subtle, on-the-fly adjustments—without having to manually switch anything on.
Watch Buddy
Watch Buddy
Watch Buddy
I also built a proof-of-concept GPT called “Watch Buddy” to mimic the conversational AI I envisioned. Stream Buddy served as a proxy for testing scenarios where a user “tells” the system about moods or disliked tropes. By observing how testers interacted with Stream Buddy, I tracked how comfortable they felt giving detailed feedback, whether it produced relevant suggestions.
I also built a proof-of-concept GPT called “Watch Buddy” to mimic the conversational AI I envisioned. Stream Buddy served as a proxy for testing scenarios where a user “tells” the system about moods or disliked tropes. By observing how testers interacted with Stream Buddy, I tracked how comfortable they felt giving detailed feedback, whether it produced relevant suggestions.
Ongoing Refinements: As testers tried summarizing their tastes or disliked tropes, I iterated Stream Buddy’s instructions to handle increasingly nuanced inputs. Each user feedback round reshaped how we structured the AI prompts and clarified the need for guided dialogue cues.
Ongoing Refinements: As testers tried summarizing their tastes or disliked tropes, I iterated Stream Buddy’s instructions to handle increasingly nuanced inputs. Each user feedback round reshaped how we structured the AI prompts and clarified the need for guided dialogue cues.
This early experiment illustrated the power of a dynamic conversation loop, which informed our eventual user journey flow, particularly the part where watchers can specify why they reject or endorse certain content.
This early experiment illustrated the power of a dynamic conversation loop, which informed our eventual user journey flow, particularly the part where watchers can specify why they reject or endorse certain content.
Try Stream Buddy!
Try Stream Buddy!
Proposed Holistic UX Approach
Proposed Holistic UX Approach
Proposed Holistic UX Approach
Unified User Data:
All watch history and preference data unify under a single user ID, ensuring consistent personalization across TV, phone, laptop, etc.
Unified User Data:
All watch history and preference data unify under a single user ID, ensuring consistent personalization across TV, phone, laptop, etc.
Simplicity-First Layout:
The default interface remains minimal, letting watchers quickly pick content. Advanced recommendation features or social commentary can be toggled on by curious or power users.
Simplicity-First Layout:
The default interface remains minimal, letting watchers quickly pick content. Advanced recommendation features or social commentary can be toggled on by curious or power users.
Detailed Preference Feedback:
Beyond simple like/dislike, watchers can specify “Too slow” or “Didn’t like comedic violence.” This knowledge fuels a more accurate rec engine.
Detailed Preference Feedback:
Beyond simple like/dislike, watchers can specify “Too slow” or “Didn’t like comedic violence.” This knowledge fuels a more accurate rec engine.
Streamlined Remote Design:
To address the cluttered nature of the current Roku remote, I designed a new, intuitive remote control. 3D modeled in Rhino, then 3D printed and painted, this redesigned remote features a simplified, ergonomic layout that enhances usability while delivering a sleek, modern aesthetic.
Streamlined Remote Design:
To address the cluttered nature of the current Roku remote, I designed a new, intuitive remote control. 3D modeled in Rhino, then 3D printed and painted, this redesigned remote features a simplified, ergonomic layout that enhances usability while delivering a sleek, modern aesthetic.
Validation & Gaps
Validation & Gaps
Validation & Gaps
Usability Tests:
Younger watchers loved aggregator features; older participants wanted simpler, direct browsing. Toggling advanced modes satisfied both.
Using the GPT-based Stream Buddy prototype to simulate conversational AI, testers described their tastes in natural language. The prototype refined recommendations accordingly, and iterative feedback helped refine prompt strategies.
This process revealed:
This process revealed:
This process revealed:
Prototype Walkthrough
Prototype Walkthrough
Prototype Walkthrough
Reflection
Reflection
Reflection
