CASE STUDY
Connecting teams, experiences, and the stories we watch.
My role
Senior Product Designer
Timeline
January 2026 — Present
Focus
Systems, discovery & strategy
Platforms
Web · Mobile · Fire TV
00
Overview
At Prime Video, I worked across people, product, process, and strategy to turn fragmented experiences into a shared design language. Not just a cleaner interface, a stronger foundation for how teams collaborate, how customers discover, and how the business grows.
01
People
Collaborating and contributing to design excellence
Situation
The fragmentation wasn’t just technical, it was a people problem. Every team had their own rules because there was no shared language, no shared standards and no culture of feedback. I needed to change how people worked together, not just what they were building.
The result was visual fragmentation across every customer-facing journey: corner radius, gutter spacing and sizing were all inconsistent. That made our system impossible to scale, govern or automate.
Action
I mapped every touchpoint I had with the people around me.
From juniors I mentored regularly to leadership I aligned with weekly, then built a rhythm that gave everyone a voice in how we raised the bar.
A rhythm for the team
Daily
Check-ins
Weekly
1:1s with junior and midweight designers, mentoring, leadership call
Bi-weekly
Design crits and design sprints
Monthly
Town halls, retrospectives, learning and development
Across regional and global leadership, directors and seniors, engineers and developers, and brand, marketing, research and legal partners.
Governance co-created, not handed down.
Governance only works if people trust it. I ran collaborative sessions to define what ownership meant, and made critique a regular practice so standards were constantly tested and refined.
Design System Governance Framework
01 / DIRECTION
Strategy and policy
02 / CONSISTENCY
Standards and guidelines
Ownership
03 / SHARED KNOWLEDGE
Documentation
04 / CONTINUOUS IMPROVEMENT
Metrics
Result
Consistency
From an estimated 250+ components across four regional libraries to one governed library of 47. Every designer on the team knew which to use, and why.
250
47
Efficiency
A 43% reduction without losing a single meaningful experience. Fewer options meant faster, more confident decisions.
28
16
Adoption, phase by phase
🇺🇸 United States
Phase 1 · Pilot
🇨🇷 Costa Rica
Phase 2 · GSS region
🇮🇳 India
Phase 3 · Production
🇬🇧 United Kingdom
Phase 4 · Ongoing coverage
02
Product
Customer insight, product quality and vision
Situation
Customer research flagged that 23% of users were failing to engage with title cards because tap targets were inconsistent. The design team received a brief with screenshots of the interface issues customers were hitting, and our job was to review, confirm, test and implement a fix.
Can we build one unified container system that works across every device, every size and every placement, and make it automatable?
One modular system.
Every device. Every size.
Action
I defined the three container ratios by auditing every existing placement and mapping them to the content types already in the ecosystem.
2:3
Portrait · Film posters and key art
A tier-based scaling system.
Consistent corner radius and gutter rules across L, M, S and XS, at every breakpoint, on every device. This is the simplified token table I handed to engineering.
One system. Every placement.
L · 1280px+
Storefront · Square
Square
Gutter: 30px Corner radius: 60px
Homepage · Square
Square
Gutter: 25px Corner radius: 50px
Schematic previews. Gutter and radius labels show production values.
From weeks of sourcing to moments: meet Figbot.
With container ratios locked, we leaned into AI-generated key art variants using Figbot, an internal Figma plugin built off this system. A designer types in a title, it connects to the Prime Video catalog and instantly surfaces the full library, correctly sized and ready to use. The team went from 48 variants per campaign to 3, which let us A/B test key art at scale for the first time.
Result
One system.
34 touchpoints.
A customer who discovers Luther in a push notification, taps through on mobile and subscribes on Fire TV now meets one coherent visual language at every step.
03
Process
Focus on methods, standards, and application
Situation
The journey showed us where to look. Browse was the broken stage.
I facilitated a design sprint with product, research, and engineering to map the full browse-to-play journey and surface where we were losing users.
Two friction points: Navigation that didn't match how users think, and ambiguous paywall signposting at the point of decision.
THE BROWSE-TO-PLAY JOURNEY
The intent was to watch. The effort was to find.
Two moments of uncertainty stood between opening Prime Video and pressing play.
01
ARRIVE
Open Prime Video
The customer lands on the home screen and sees featured titles and autoplay promotions.
Curious · Ready to discover
02
DISCOVER
“Where is something I want to watch?”
Scroll carousels, explore categories or search for a title.
Hopeful, then overwhelmed
FRICTION 01
Navigation doesn’t match the customer’s mental model.
Categories and recommendations make it hard to browse by mood, occasion or format.
DROP-OFF RISK
Keep scrolling, return to a familiar title, or leave without choosing.
Design opportunity: Organise discovery around how customers think.
03
EVALUATE
Open a title and decide
Read the synopsis, check ratings and watch a trailer to decide whether it is worth watching.
Interested · Looking for reassurance
04
COMMIT
“Is this included, or do I pay extra?”
Check whether the title is included with Prime or requires an add-on subscription, rental or purchase.
Uncertain · Hesitant to commit
FRICTION 02
The paywall becomes clear too late.
Ambiguous signposting interrupts the decision and breaks the customer’s context.
DROP-OFF RISK
Back out, restart the search, or leave before playback.
Design opportunity: Make inclusion and payment requirements clear before the decision.
05
PLAY
Choose confidently. Start watching.
Select the title and begin playback. Adjust subtitles, audio and quality as needed.
The intended destination · Engaged
Action
Alex wasn’t disengaged. He was overwhelmed.
The research team and I ran eight user interviews and synthesised findings into empathy maps to reframe the problem. This wasn’t a disengagement issue; it was a navigation mismatch. Alex grouped things by mood, occasion, and format. The app organised by editorial taxonomy. That mismatch showed up in every quadrant.
02 / A GUIDED EMPATHY STORY
01 — HEARS
The evening starts with a recommendation.
“You have to watch The Boys. It’s on Prime!”
Alex arrives ready to watch. Friends have already done the convincing. His task should be simple: find it, check it’s included, press play.
THE SIGNAL / Motivation isn’t the missing ingredient.
We stopped assuming. We started asking.
Working alongside the user research team, we ran a hybrid card sort study with 20 participants using UserTesting. We chose this method because it gave us both structure and the freedom to surface unexpected mental models. The study used 35 cards and six starting categories, with an estimated completion time of 8–12 minutes.
Users knew exactly where some things belonged. Others, not at all.
The heat map told a clear story: Peppa Pig, Oppenheimer and Premier League had near-unanimous agreement. Users placed these instinctively. But these were also the zones with the highest browse abandonment. The navigation wasn’t broken everywhere; it was broken in the places that mattered most.
Result
Three methods. One clear direction.
The journey map told us where to look: browse was long but low-converting, so we made it the primary intervention target. The empathy map told us why: cognitive load, not disinterest, was driving drop-off. That set the path for card sorting and A/B testing. The card sort told us what to change: 11 of 14 categories couldn’t be placed confidently, especially Prime Originals versus channels. We collapsed categories from 14 to eight, opening discovery work on Prime Originals. The findings are now informing a live A/B test on navigation labels. This remains a work in progress.
04
Position
Focus on strategic influence and organisational embedding
Situation
A business that wanted to be two things at once.
Subscriptions keep Prime Video financially viable, with 30 channels per country generating recurring revenue. At the same time, the strategy was to elevate Prime Originals in the global streaming market, despite a catalogue of only 160 original titles. Research showed customers couldn’t distinguish Originals from subscription content and assumed everything was included with Prime. We had a differentiation problem, not just a content-volume problem. How could we visually differentiate the two product types without undermining the revenue model that funds the business?
User sees
“I get Prime Video for free with my Amazon subscription.”
No distinction between Originals and subscriptions.
Business says
Keep subscriptions prominent. They pay the bills.
Strategy says
Elevate Originals and compete in the streaming market.
Action
Strategy versus business goals.
I reframed the strategy conversation with content-gap data. Prime Video had 3.6% of Netflix’s output, shifting the conversation from leading with Originals to asking how we could do both. For the VP of Product, revenue protection meant subscriptions retained prominence while Originals were elevated alongside them. For the Brand Director, Originals received a distinct visual language for the first time.
Originals output across streaming platforms
4,400 titles
300 titles
260 titles
160 titles
3.6% of Netflix’s output. How could we do both?
We gave Prime Originals a visual identity.
A new Prime Original logo, a premiere countdown and a distinct treatment in navigation established an identity alongside subscriptions. Characters and context were revealed while keeping the mystery alive.
Two entry points. One consistent identity.
A dedicated Prime Originals banner reinforces the identity for customers who reach the storefront without going through navigation. Prime Blue hover states highlight Original content in context across the interface, making the distinction clear without disrupting the browse experience.
Result
Design changed the timeline.
We deliberately chose not to bury subscriptions in favour of Originals: the business couldn’t afford to lose channel revenue. The design answer was additive, elevating Originals with a distinct identity without deprioritising what funds the platform. The phased roadmap replaced an immediate Originals-first plan and was agreed at VP level. Distinct badging, colour language and card hierarchy shipped across web and app. Product, Brand, Engineering and Research aligned on a shared visual framework for the first time.