Melrose Trading Post • Website Redesign

Fixing a flea market that forgot its users


Role

UI/UX Designer

Context

Bootcamp Capstone

Timeline

2 Weeks

Tools

Sketch • Figma

01 - Context

A beloved market with a website that worked against it

Melrose Trading Post is a weekly outdoor market and event venue in West Hollywood — vintage finds, live music, local vendors. The website told a very different story

The site had chaotic navigation, an event-booking flow that more than half of users couldn't complete, and brand visuals that didn't communicate the energy of the place. Ticket sales for performances were the direct casualty.

52%

Of 25 survey participants failed to add a ticket to cart

7→4

Primary nav items reduced from 7 to 4 clear categories

9→5

Total navigation categories collapsed via card sorting

2 weeks

Full process research > synthesis > IA > Prototyping > Testing

02 - Research

I didn't guess at the problems. Heuristic evaluation, a quantitative survey, and user interviews ran in parallel to build a multi-layered picture of where the site was breaking down.

Three methods. One clear diagnosis

Heuristic Evaluation - Nielsen’s 10 Principles

I evaluated every screen of the current site against Nielsen's heuristics, rating each issue by severity. Three Severity 4 issues emerged immediately — catastrophic problems that would block almost any user from completing their primary goal.

The through-line: the site was designed around the organization's structure, not the user's mental model. Categories like "Participate" made sense internally but were opaque to someone who just wanted to attend a show.

4

Catastrophic before launch

  • No clear way to navigate the site from the home page

  • Events can only be booked through one single confusing path

  • Users cannot backtrack from a non-existent event page

3

Major - High Priority

  • Home page doesn't clearly show what events are available

  • It takes excessive effort to find and access events

2

Minor - Fix if time allows

  • No shortcuts or accelerators for returning users who know what they want

1

Cosmetic - Fix last

  • Visual inconsistency in typography and brand treatment across pages

Online Survey - 52% failure rate

25 people were asked to navigate the current site and add an event to their cart. This is the site's primary user goal. Just over half couldn't do it.

That single number justified the entire redesign effort. This wasn't a cosmetic problem — it was a structural one.

Card Sorting - 9 categories were brought down to 5

Open sorting let users group the existing navigation however felt natural. Without prompting, they collapsed 9 categories into 5. Closed sorting confirmed the groupings and validated the new labels.

Key insight: "Participate" confused users who thought it meant becoming a vendor. Renaming it "Get Involved" eliminated the ambiguity entirely.

03 - Define

The site map was the diagnosis and the prescription

Mapping the existing site structure revealed the problem at a glance: "Participate" had 8 sub-pages — more than all other sections combined — while the content users actually came for was buried or missing.

Current sitemap

Redesigned Sitemap

Current vs. Redesigned User Flow

I mapped the existing booking flow step by step, then designed the future-state flow. The comparison made the problem concrete: users in the current flow had to navigate between two separate sites (MTP and Greenway Theatre) with no handoff, and could end up on dead event pages with no way back.

The redesigned flow collapses this into a single continuous path with clear breadcrumbs at every step.

❌ Current Flow

  • Home page - no clear event. CTA above fold

  • Search bar (non-obvious, often skipped)

  • Calendar is buried in a secondary navigation

  • Dead event page - no backtrack option

  • Jump to separate Greenway site to buy tickets

  • 52% abandon rate before completing a transaction

☑️ Redesign

Redesigned the iconography across all menus.

  • Hero immediately features next even + Get Tickets

  • Display the event calendar as a primary nav item.

  • Update event cards with date, image, and event info button.

  • Select performance will make it clear which show the user is selecting

  • Ticket stepper -Adult/Child, live subtotal.

  • Finish checkout - breadcrumb exit is always visible.

04 - Design Process

From empathy to prototype in 2 weeks

01 - Empathize

Heuristic evaluation, 25-person survey, user interviews to map mental models

02 - Define

Card sorting, current site map audit, journey mapping, persona creation

03 - Ideate

Feature prioritization, revised site map, IA restructure from 9 categories to 5

04 - Design

Paper prototype → mid-fi wireframes → hi-fi in Figma with full visual system

05 - Test

Usability testing on prototype, iterate on checkout flow and event card layout

05 - Mid to Hi Fidelity Prototype

The final prototype introduced a consistent visual system — teal nav, legible event cards, and a checkout flow a first-time user could complete without help.

The Redesign screen by screen

Mid-fidelity home page

Mid-fidelity Calendar page

06 - Design Desicions

The Tradeoffs I worked through

Every decision had an alternative. These are the choices that shaped the final design and what I considered but ruled out.

07 - Reflection

What I learned and what’s still unfinished

What real data gave me

The 52% failure rate wasn't a gut feeling — it was a number. That data gave me moral authority to make bold structural changes that would have felt presumptuous without evidence. Research doesn't just inform design; it justifies it.

Whats genuinely unresolved

Mobile. I designed at desktop width, but MTP's audience discovers events on their phones. The event cards, ticket stepper, and checkout form all need rethinking for a 375px viewport. That's the biggest gap in this work.

What I’d do differently

I'd run a second usability test specifically on the checkout flow — ticket quantities, the Back button, and what happens on a sold-out show. I designed for the happy path. Edge cases need testing too.

What I’d validate before shipping

"Event Calendar" tested well in card sorting — but card sorting and real-world use are different things. I'd run 5 first-click tests on the final nav before considering this shipped.

08 - Rationale

Key Design Decisions

09 - Next Steps

Where this goes from here

The hi-fi prototype proved the concept. Three clear development paths were identified for the next phase.

01 - Partner with car manufacturers

Build white-label versions for individual brands — incorporating their model libraries, color systems, and pricing alongside Divergent 3D's printed components.The "rescan" button caused confusion about its purpose and timing

02 - Integrate dealership inventory

Connect to live inventory data so the app surfaces real availability, actual dealer pricing, and nearby locations for the exact configured vehicle.

03 - Highlight Divergent 3D parts in context

Build a dedicated view that identifies exactly which components are Divergent-printed — making the company's core technology visible and meaningful at the moment of decision.

10 - Reflection

What this project taught me

Testing early is always worth it

The most valuable insight from this project didn't come from the hi-fi prototype — it came from the first paper session, when users were confused by the "rescan" button and couldn't navigate back to home. That finding shaped every subsequent version at zero visual design cost.

In a two-week sprint especially, the instinct to skip early testing in favor of getting to "real" screens is exactly backwards. Low fidelity gives you permission to be wrong when being wrong is cheapest

AR interaction design needs its own playbook

Most mobile UX conventions assume a 2D screen with a fixed viewport. AR breaks both assumptions — the environment is 3D, the viewport moves, and the "screen" is the real world. Patterns that work perfectly in a standard app can actively undermine an AR experience.

Given more time, I'd have explored how AR-native patterns — spatial anchoring, gaze-triggered menus, gesture controls — might reduce reliance on on-screen UI altogether, making the experience feel less like an app and more like a showroom.