MARKET
SQUARE.
A peer-to-peer marketplace for high-value goods where trust is the product. I rebuilt onboarding, listing, and verification so the experience earned confidence instead of asking for it.
A peer-to-peer marketplace for high-value goods where trust is the product. I rebuilt onboarding, listing, and verification so the experience earned confidence instead of asking for it.
MarketSquare is a U.S. fintech building infrastructure for large, time-sensitive payments. Their next bet was an iOS marketplace for high-value peer-to-peer transactions — and the web product had already proven the model worked.
The web experience was functionally complete, but conversion was leaking at every seam. 64.3% of users abandoned before listing their first item. Sellers couldn't tell whether their listings were clear. And because buyers and sellers had to make contact to move a deal forward, every interaction carried the quiet question, "is this a scam?"
The diagnosis underneath all three symptoms was the same: friction was eroding trust, and without trust there is no transaction confidence. In a marketplace, that compounds — low trust suppresses liquidity, and thin liquidity makes the next user trust you even less.
More friction → zero trust → no transaction confidence. Fix the friction and the trust follows.
I framed the engagement around three outcomes the business and its users both needed — not three features to ship, but three behaviors to change.
Reduce friction and cognitive load through onboarding so first-run never feels like a form.
Introduce AI where it measurably improves listing clarity — not as a headline feature.
Reduce churn by building verification and safety into the flow, so trust is structural.
I opened with a kickoff to align the team on goals, scope, and constraints, then co-facilitated a workshop to map pain points and pressure-test whether the journey needed simplifying or rebuilding.
The hard part of a two-sided marketplace is that the two sides are in tension. Unlike linear e-commerce, every lever that reduces friction for sellers can introduce risk for buyers, and vice versa. Liquidity depends on keeping both sides moving at once — so the work wasn't to optimize a funnel, it was to hold a balance.
Build durable trust in the product, and use AI to help sellers evaluate and list items fast enough to keep inventory flowing.
A shorter path in, listings that are clear to read and easy to fix, and protection from scams without surveillance-grade friction.
How might we mitigate scam risk, streamline onboarding, and introduce AI for in-app clarity — without trading one side of the marketplace against the other?
I audited prior and current research rather than restarting it — fast teams earn speed by reusing evidence. The data converged hard: fatigue concentrated in three moments of the flow.
Every interest category got its own screen, with no way to skip. Users cycled 10+ frames before they ever saw the home feed.
Sellers couldn't edit a listing once posted — no price or description updates. One mistake meant a dead listing.
Users hesitated to share personal and financial details with nothing signalling the other party was real.
To make the breakdowns legible to the team, I rebuilt the user flow to mark exactly where friction and trust collapsed, then layered a journey map from real usage data — splitting the experience by the two roles a single person plays: buyer and seller.
With tight deadlines, I treated iteration as the decision-making tool. I moved fast from lo-fi wireframes into hi-fi as choices firmed up, keeping product and engineering in the loop so every refinement was a shared call, not a reveal.
Stakeholders arrived with a strong picture: a defined dock with a matching header, and a home feed of auto-playing looping videos — an explore page meant to out-flash the competition. It was a reasonable instinct, and the wrong one for this product.
I made the case for less. We were trying to reduce cognitive load; auto-playing video and a two-toned chrome added motion and decision-cost in exactly the moment we needed calm. I proposed a clean light/dark system in place of the two-tone concept, and reframed the goal: the home feed's job is to make the next tap obvious, not to entertain. The team came along.
That set up the real choice — Iteration 2 vs 3. Two was the obvious pick: it fit everything on one screen without scrolling — logo, search, filter, categories, distance, and six full products. I argued for three. Its larger image surface gave products room to sell themselves, and folding distance into the filter icon removed a permanent control most users touched once. We shipped three.
I built light and dark prototypes and ran them A/B against the existing experience — so the redesign had to win on evidence, not taste.
Feedback was direct and useful. It confirmed the shortened onboarding landed, surfaced where verification reassured rather than annoyed, and flagged the one feature that helped one metric while quietly costing another.
Buyers and sellers felt they could navigate the new design with more ease compared to previous iterations.
The introduction of a “skip” button during onboarding increased user satisfaction.
Overall positive. Dark mode preferred by 70% of the test group.
Reducing the number of onboarding screens drove a 20% reduction in abandonment.
The brand stood for luxury, royalty, and high-value — qualities that tempt a designer toward heavy color. I went the other way.
I built the palette mostly from variations of black and white — generous negative space carrying the "high-value" feel — and reserved the brand purple for accent and intent. That choice did double duty: it held the premium tone and kept the product compliant with WCAG contrast across both light and dark modes. Restraint was the luxury.
At this pace the language couldn't live in screens alone — it had to live in a system. I built MarketSquare's first design system: color, type, spacing, and radius tokens feeding a component library of buttons, fields, chips, listing cards, the navigation dock, and the verification patterns that carried trust through the product.
Everything was documented in light and dark with WCAG-checked contrast, built on an 8-pt grid with Figma variables and auto-layout, and handed off dev-ready. It cut design debt, kept the team consistent while we shipped fast, and gave the web product a foundation to inherit.
How might we reduce friction and cognitive load during onboarding?
Onboarding used to give every interest category its own screen with no escape hatch — 10+ frames before the home feed. I condensed categories into a single selectable view and added a Skip, so committing to the product no longer required completing a survey first.
How might we introduce AI to improve listing clarity?
Sellers previously couldn't edit a listing after posting — no price or copy fixes. I introduced AI-assisted listing that drafts a clean summary and suggests tags, paired with full edit control after publish. AI does the first draft; the seller stays in command.
How might we reduce churn and increase trust in the product?
I built safety into the structure: MFA and ID verification so both sides knew the other was real, plus a pay-to-talk gate on inquiries. The verification work made buyers and sellers measurably safer — and pay-to-talk filtered for serious buyers, a deliberate tradeoff I flagged because it lifted intent while dampening raw engagement.
Lower cognitive load, higher satisfaction, and stronger conversion — and a verified, AI-assisted foundation the team could now scale across the product, including the web version it started from.
In a two-sided marketplace, trust isn't a feature you add — it's the equilibrium you maintain. Every decision had to serve both sides at once, or it served neither.
Optimizing for the fewest steps did the heaviest lifting. Condensing onboarding and adding a skip moved satisfaction more than any single screen redesign.
Translate the AI-assisted, verification-first model to web — and extend MFA into the wallet flow, where the trust stakes are highest.
Open to full-time roles, freelance projects, and conversations about design that moves the needle.