← Back to work
Case Study 01

Virtual Closet

2025 – 2026 · UX Designer · Figma

Live Prototype

See it in motion.

The screen on the right is a snapshot of the actual working prototype — open it full-screen to browse the OOTD recommendations, check the Weekly Style Planner, and try the sustainability score for yourself.

Open full-screen
Virtual Closet home dashboard: Good Morning Emma weather card, OOTD Recommendations, Weekly Style Planner, and Recently Worn
01 · Overview

A mobile ecosystem connecting active wardrobe management, contextual outfit planning, and friction-free shopping in one place.

Role
UX Designer
Timeline
Jun–Sept 2025, Apr–Jun 2026
Tools
Figma

The project paused between those two working periods for a CSL Behring graphic design internship, then resumed in April 2026.

02 · The Insight

Users regularly wear fewer than half the items in their closets. That gap drives daily decision fatigue and outfit-coordination anxiety.

03 · Structure & Flow

Dashboard split into two zones — Upload Clothes and Get Outfit Suggestions — to separate passive wardrobe management from active decision-making. Added an Outfit of the Day module using live weather data so suggestions felt contextually relevant rather than generic.

04 · The Scope Decision

Early plans included a 3D Virtual Try-On avatar simulator. I parked it deliberately (kept as a future-tier placeholder) to protect quality on three core pillars: closet logging, outfit matching, and sizing trust.

05 · Key Decisions
Decision: Onboarding
Problem
Multi-stage "What's your style?" quiz blocked entry.
Evidence
2 of 5 usability participants abandoned the flow.
Decision
Trimmed to a single prompt; preferences moved to a progressive Profile tab.
Removing choice architecture from the critical path let users reach value immediately, deferring personalization to a moment they'd already opted in.
Decision: Outfit Suggestions
Problem
12-item grid with overlapping tags caused choice paralysis.
Evidence
No clear signal for why any match was prioritized.
Decision
Single focal card with a "95% Match" score; secondary logic behind a "Why this works" accordion.
A single ranked recommendation reduces the decision from "which of 12" to "yes or show me why." Complexity is still available, just not mandatory.
06 · Trust Microcopy

Sizing recommendations are grounded in the user's own closet brands (not abstract percentages), reframing an ML calculation into personal, trustworthy advice.

07 · Impact
Onboarding Completion
2/5 → 100%
Friction Points
11 → 5
Core Pillars
3/3 Verified
08 · Retrospective

Learned to ruthlessly prioritize fixes by impact (11 friction points trimmed to 5), document design rationale in real time, over-recruit for usability sessions, and treat intentional omission as a design skill, not a shortfall.