Hero — Final Design Outcome

How do you design software that moves $79B worth of assets daily? My design turned a risky, tedious, error-prone ETF portfolio management task into exception-driven experience. Instead of 12,000 lines of data, users now view 560 flagged ones that truly need their attention.

The Trust Problem

Trust isn’t binary — users don’t flip from “I don’t trust this” to “I trust this.” They build confidence through micro-verifications, and the system’s job is to support that progression at every layer.

Before: Daily workflow with multi-loop vendor correction cycleImage 1 of 4
After: Streamlined Meteor workflow — auto-generate, flag, review, confirmImage 2 of 4
12,000 → 560: Visual noise reduction comparisonImage 3 of 4
Sanitized screenshot of basket review interface showing flagged vs. unflagged linesImage 4 of 4

12,000 lines → 560 flagged · Review time: 80 min → 8 min per basket · 45+ baskets daily

Leverage-Based Scoping

Upstream errors cascade — if basket management is wrong, every downstream order is wrong too. I mapped the full portfolio management lifecycle to find where one designer’s effort would have the highest leverage.

ETF Portfolio Management Cycle: Fund Launch → Holdings → Basket → OrderImage 1 of 3
Coverage matrix: FI/EQ × lifecycle stages with existing tools mappedImage 2 of 3
Upstream → Downstream funnel: basket management as the highest-leverage bottleneckImage 3 of 3

ETF lifecycle: Fund Launch → Holdings → Basket Management → Order Management

Adoption Sequencing

Same product type doesn’t mean same workflow — Fixed Income and Equities have fundamentally different tool readiness, mental models, and failure modes. I chose the team whose pain was acute and switching cost low, knowing an early win would unlock adoption across the floor.

A Tale of Two Teams: EQ vs. FI adoption readiness comparisonImage 1 of 3
User research insights: side-by-side EQ (desperate for change) vs. FI (entrenched habits)Image 2 of 3
Scoping matrix with EQ 1st Priority / FI 2nd Priority annotationsImage 3 of 3

EQ: zero internal tooling, desperate for change · FI: legacy tools, entrenched habits · EQ first → internal precedent

ETRO — Progressive Trust Calibration

I designed a progressive trust calibration framework — four principles that scaffold human confidence in automated data: Explainability (show me why), Traceability (show me what changed), Reversibility (let me undo), Observability (make me accountable).

Explainability: Severity tier reasoning — flagged row with reasoning tag + decision treeImage 1 of 5
Traceability: Corporate actions diff view — before state, after state, deltaImage 2 of 5
Traceability detail: Corner notch indicator showing what changed and by how muchImage 3 of 5
Reversibility: Pro-rata calculation sandbox with preview before committingImage 4 of 5
Observability: Override justification flow — deliberate friction for audit trailImage 5 of 5

Severity tier reasoning · Corporate actions diff view · Rebalancing preview sandbox · Override justification with deliberate friction