
Creating a leaderboard system that brings clarity across content types, and a destination worth returning to for the best content.
TIMELINE
2024–2026
ROLE
Lead Designer
DELIVERABLE
Components + vision
TEAM
Product Management Data Science Engineering Design Systems
Ranking & recommendations
Content discovery
Product strategy
Systems thinking
Design systems
01
/
10
Charts · 2024–2026
Making Charts readable at a glance.

Charts as designed: leaders on top, then numbered carousels, ranked lists and category tiles.
02
/
10
The problem
Every row looked the same.
Problem
Discover had no clear purpose and was set to close. It was renamed Charts with rows of "Top ___", but visitors still did not understand what the page was for.
Opportunity
Decide the page's job, then make the design say it: rank, hierarchy, and filters that scale.
Insight
Every row looked the same, so nothing signaled which mattered most. The page looked like a recommendation feed but behaved like a ranked leaderboard.
Objective
Make Charts read as a leaderboard that shows creators what the platform wants to encourage, with player value as a constraint.

Charts before: rows of “Top ___” on tablet and phone.
03
/
10
Role and approach
Lead designer for the surface.
Role
Rank, hierarchy, filters, and the reusable components behind them.
Team
Product, data science, engineering and design-systems partners.
Alignment
Weekly design reviews, biweekly product reviews and a quarterly leadership presentation. I phased the vision so product and engineering could commit one step at a time.
Approach
Settle the page's job, design the system around it, ship what the team could, and design the rest ahead with a staged rollout plan.

The surface as a system: leaders, carousels of five and ten, ranked lists, categories.
04
/
10
The job
Charts work like a leaderboard.
Where we landed
Charts is a leaderboard. Personalized browsing belongs on the search landing page.
Why creators
The vision was a deterministic, non-personalized leaderboard for creators, with player value as a constraint.
Personalized browsing became its own case study:

Trending this week, ranked one to ten.
05
/
10
Chapter A · Show rank
If the titles are numbered, the order explains itself.
A wide range of numbering explorations for grid tiles and list items: type styles, alignments, and digit widths of 1, 44 and 100. Each was pushed across tile sizes, aspect ratios, device types and screen sizes, looking for one solution that would scale.

Numbering explorations at 1, 44 and 100, on tiles and list items.
06
/
10
Chapter B · Create hierarchy
Importance is visible before you read a word.
Oversized numerals on top, stepping down to numbered carousels of five and ten, then ranked lists and category tiles.

System logic: scale decreases as content density increases.
07
/
10
Chapter C · Narrow the chart
Narrow the chart without leaving it.
Chips for one quick change, a menu for many. Every selection stays on screen as a chip, and its × removes it, so you can adjust without reopening the menu. The components were adapted into the design system for use across surfaces.

Closed, chips for one quick change stay above the chart. Open, select once and adjust in place.
One multi-filter menu · collapsed and expanded

08
/
10
Outcome
Where the work landed.
Delivered
Row logic, ordering and quality, genre and music rows, an info icon, and the first filters as basic chips and dropdowns.
In the design system
The responsive filter menu, built from the new components. Fuller versions went into creator tools.
In build
Genre pills were built; sticky filters were in development.
Designed ahead
Hierarchy, trending indicators and the consolidated filter panel, with numbering at proof of concept.
09
/
10
Takeaways
What this work shows.
Design the whole system
Rank, hierarchy and filters were designed together, so any piece could ship first.
Stage the vision
Start with the smallest visible change, a number on a tile, and land the rest in steps.
Interpret executive input
I treated mandates as inputs to design around.
10
/
10
What I learned
What I learned.
Systems and staging
Systems, staging, and prioritizing: how to break a holistic vision into steps a team could actually take.
Give the vision an identity
The value of giving a vision a real identity: a designed picture of what the surface could be, so people believe in the whole thing.
Test the hierarchy earlier
I’d test the hierarchy earlier with creators and players separately, to confirm what each needs from the leaderboard.