The Problem

A luxury brand trapped in Instagram DMs

A luxury artisan chocolatier was operating entirely through Instagram DMs and phone calls. No website. No checkout. No trust signals. Manual order processing was creating bottlenecks, losing customers, and making international expansion difficult.

Cultural skepticism compounded the challenge: a meaningful share of prospective Kazakh and Russian customers told the client they wouldn't purchase without a professional, dedicated e-commerce presence. Cart abandonment on existing informal channels was estimated by the client at around 78%. The brand had a premium product and no digital infrastructure to support it.

Before · Instagram DMs

DMs and phone calls only — high friction, no scale

No payment flow, no trust indicators, no analytics

An estimated 78% cart abandonment on informal channels

After · Final Design Delivered (Not Yet Built)

Trilingual UI — English, Russian, Kazakh with locale switching

Localized payments designed in — Kaspi, QIWI, Visa, Mastercard

87% task completion, 82 SUS score, validated in real usability testing

Design Screens

Figma Concept Screens — Prototype Only

Every screen below is from the final high-fidelity Figma prototype, tested with 15 participants across Kazakh, Russian, and English-speaking markets. These are concept designs — not screenshots of a live website. No such site was ever built.

Process Breakdown

Three phases of real work. Two honestly marked as plans.

Every phase was tightly scoped and evidence-gated. Here's how the project actually ran — including being direct about where the work stopped.

What I Learned

Lessons that shape every project since

User Feedback

What testers said

From real usability testing sessions, anonymized since participants did not consent to being named publicly.

The Problem

"Feels like using a system from another era"

Craigslist, while iconic, still asks users to build trust infrastructure outside the product itself. Anonymous listings, no credibility signals, buried filters, and desktop-centric layouts create a friction-filled experience that rewards patience over confidence.

I discussed this problem informally with other designers and developers, and ran a short survey and 1-on-1 interviews to understand where the friction actually lived — before designing and usability testing a concept redesign with 18 real participants.

Before · Observed on the Real Product

Anonymous listings, no credibility signals

Filters buried or missing, image previews inconsistent

Desktop-centric layouts create friction on mobile

After · My Concept (Prototype Only)

Seller profiles with trust badges and credibility indicators

Persistent filter sidebar on desktop, mobile-friendly filter modal

84% task completion, 40 seconds faster trust flow in real testing

Design Screens

Figma Concept Screens — Prototype Only

Every screen below is from my Figma prototype, tested with 18 real participants before a single line of production code would have been written. These are concept designs — not screenshots of a live Craigslist product.

Figma Prototype

Visual browsing concept with advanced filters for faster product discovery.

Figma Prototype

Verified seller profile concept showcasing trust signals and activity history.

Figma Prototype

Search and filter controls concept for finding relevant listings quickly.

Listing view concept with a reporting flow for suspicious listings.

Figma Prototype

Image-first catalog layout concept designed for effortless comparison and browsing.

Figma Prototype

Posting flow concept for creating and managing a listing.

Figma Prototype

Filtered results view concept that reduces clutter and improves decision-making.

Seller credibility concept integrated directly into the listing view.

Process Breakdown

Discovery → Design → Validate → What I'd Build Next

The first three phases below are real work I completed and tested. The last two are honest, detailed plans for what I'd do next if this moved into development — written to show exactly how I think about implementation and launch, not to claim they happened.

01 · Discovery 02 · Design System 03 · Validation 04 · My Implementation Plan 05 · My Launch Plan

Phase 01 · Discovery

Mapping frustration to specific interaction failures

I discussed this concept informally with other designers and developers, and ran a survey plus 1-on-1 interviews to understand how people currently build trust with an anonymous seller before ever messaging them. The goal wasn't general dissatisfaction — it was specific failure moments.

I also reviewed several existing marketplace apps for patterns — looking at how each handled seller verification, filtering, and mobile posting differently, to ground my design in realistic conventions rather than invented ones.

Key theme: Participants described doing external research — cross-referencing phone numbers, checking for scam patterns — before ever messaging a seller. The product was pushing users to build trust infrastructure outside of it, not inside it.

18

Usability test participants

Junior designers and everyday Craigslist users, recruited via a design community Slack, tested remotely over video.

Several

Marketplace apps reviewed

Informal review of existing marketplace patterns for trust signals, filtering, and mobile posting flows.

Survey +
1:1s

Informal research methods

Discussed the concept with designers and developers, and ran a short survey and 1-on-1 interviews before designing.

Phase 02 · Design

"Calm and competent" — not flashy

The design direction was deliberately restrained: neutral fonts, controlled palette, trust-first layout hierarchy. The goal was a system that felt like a public utility the community already owned — not a startup rebrand.

Palette: charcoal (#2A2A27), warm off-white (#F7F5F0), teal accent (#0D7A5F) — the same system I use across my own portfolio, chosen for calm legibility over visual noise.

1

Listing card component

Designed grid and list variants sharing a common data model. Trust badges surfaced at the card level rather than buried in a separate profile page.

Figma component variants

2

Filter system — persistent sidebar and mobile modal

Desktop: persistent left sidebar. Mobile: full-screen modal, tested against the original drawer pattern as part of the usability sessions.

Tested in usability sessions

3

Seller profile concept — evidence before self-description

Profile design led with credibility indicators before any self-written bio content — making verifiable signals the first thing a buyer would see.

Trust architecture

4

Reporting flow redesign

A redesigned flow for reporting a suspicious listing, tested directly with usability participants as one of the three core tasks.

Tested in usability sessions

Phase 03 · Validation

Testing the concept with real participants

I ran usability tests with 18 participants — a mix of junior designers and everyday Craigslist users, recruited through a design community Slack, tested remotely over video.

Task	Completed	Result
Post an item using the new trust flow	15 of 18	83% completion
Verify a seller's identity	17 of 18	94% completion
Report a suspicious listing	14 of 18	78% completion
Comparison: old vs. redesigned trust flow	Moderated sessions	Redesigned version 40 seconds faster on average

84%

Overall task completion

Not run

SUS score

40 sec

Faster trust flow

18

Real test participants

Key findings: Trust badges increased participant confidence in initiating contact with a seller. The reporting flow felt smoother to use. Identity-verification wording caused minor confusion — flagged as a copy fix for a future iteration.

Phase 04 · Plan, Not a Result

How I'd approach implementation

This concept never reached a development phase — there was no engineering team, no codebase, nothing to hand off to. But I designed it with a real build in mind, and here's specifically how I'd approach implementation if it did move forward.

Deliverable	Format	Why This Approach
Component specs	Figma Dev Mode + annotations	Removes redline back-and-forth — engineers pull exact values directly
Design tokens	Exported as JSON, mapped to CSS custom properties	One source of truth between Figma and code, so palette or spacing changes don't drift
Staged rollout	Internal → 10% → 50% → 100%	Surfaces edge cases — like the identity-verification wording confusion found in testing — before they reach everyone
Analytics instrumentation	Event tracking on trust-flow and reporting-flow interactions	Lets the team validate in production whether the testing gains (84% completion, 40 sec faster) hold at scale
Dev syncs	Bi-weekly, focused on the trust-flow and reporting components specifically	These were the two most complex interactions in usability testing — worth the extra review time

Why this matters to me: Design work that never reaches engineers rarely reaches users. I structure specs and handoff documentation the same way whether or not a project ships, because that discipline is what actually determines whether a design survives contact with a real codebase.

Phase 05 · Plan, Not a Result

How I'd approach launch and measurement

This was never deployed by Craigslist — there is no real launch data, and I won't invent any. What I can share honestly is exactly what I'd want to measure and why, based on what the usability testing already told me.

What I'd Track	Why
Trust-flow completion rate in production	Testing showed 84% completion in a moderated setting — real-world, unmoderated behavior is the real test
Suspicious-listing report rate before/after	Participants found the redesigned reporting flow smoother — I'd want to see if that translates to more reports actually getting filed
Identity-verification drop-off point	Testing flagged confusing wording here specifically — I'd watch this funnel closely post-launch to confirm the fix worked
Filter usage rate (desktop vs. mobile)	To validate the sidebar/modal split actually serves both contexts well, not just in testing

Honest end point: A tested, validated design concept — ready for exactly this kind of measurement plan, never actually deployed by Craigslist. I'd rather show you how I think about launch than invent a launch that didn't happen.

Lessons

What this project taught me

User Feedback

From real usability testing sessions

The Problem

Saved items on most platforms go unused

Saving something is easy on nearly every platform. Finding it again later, when it actually matters, usually isn't. LinkedIn's Save feature is a flat, chronological list with no categories, no tags, and no way to group things by why you saved them in the first place.

I wanted to explore whether structured organization and better retrieval cues could turn a passive archive into something people actually returned to — not by assuming the answer, but by designing a concept and testing it with real people.

Before · Observed Pattern

Flat chronological list — no organization, no context

No privacy signals — ambiguous who could see saved activity

Desktop-centric discovery, no reminders

After · My Concept (Prototype Only)

Smart Collections grouped by topic

Explicit 3-tier privacy control — Only Me / Connections / Public

93% task completion, 78 SUS score, validated in real usability testing

Design Screens

Figma Concept Screens — Prototype Only

The screens below are from my Figma prototype, tested with 22 real participants. They are concept designs — not screenshots of a live LinkedIn feature.

Figma Prototype

Create New Category concept: pop-up for adding a new category, with text input, color picker, and confirmation buttons.

Figma Prototype

Category Selection Grid concept: categories displayed with save counts, selected category highlighted for clarity.

Figma Prototype

Saved Post Detail concept: individual saved item view showing author, role, hashtags, and a dropdown for assigning categories.

Tag Filter Dropdown concept: tag-based filtering menu, including Product Management, UX Design, Remote Work, AI & ML.

Figma Prototype

Save Activity concept: weekly and monthly save counts visualized with progress bars, highlighting the top category.

Figma Prototype

Smart Collections concept: auto-grouped saved posts by theme, including Career Insights and #Product Management.

Figma Prototype

Quick-Access Panels concept: a curated list of pinned and recent resources, including articles and job posts.

Generate Folders from Profile concept: auto-created folders based on profile data, like past employers and skills.

Process Breakdown

Research → Design → Validate → What I'd Build Next

The first three phases below are real work I completed and tested. The last two are honest, detailed plans for what I'd do next if this moved into development — written to show exactly how I think about implementation and launch, not to claim they happened.

01 · Research 02 · Design 03 · Validation 04 · My Implementation Plan 05 · My Launch Plan

Phase 01 · Research

Why save something you'll never look at again?

I discussed this problem informally with other designers, PMs, and a LinkedIn product designer I know — not as an official engagement, but as peer feedback to sharpen the direction. The consistent theme: saving is frictionless, but retrieval is an afterthought on almost every platform, LinkedIn included.

That informal research shaped the concept direction before I moved into design and, later, real usability testing with 22 participants.

Working theory: The problem isn't that people forget what they save — it's that the product gives them nothing to remember with. No folder, no tag, no reminder. This became the core hypothesis the design and later the testing were built to validate.

22

Real usability test participants

Mid-career professionals and recruiters, recruited via LinkedIn outreach, tested asynchronously with video follow-ups.

Informal

Peer feedback

Discussed with designers, PMs, and a LinkedIn product designer — informal input, not an official engagement.

3

Core tasks tested

Save a job post, organize into folders, retrieve saved content — the three behaviors the concept needed to prove out.

Phase 02 · Design

From flat list to organized system

I designed toward three specific ideas: smart, low-effort organization; explicit privacy control; and better retrieval cues for time-sensitive saves. Working within a LinkedIn-like design language, introducing new patterns only where the existing model couldn't support the interaction.

1

Smart Collections — auto-grouping by inferred tag

Collections auto-group saves by topic (#ProductManagement, #CareerGrowth, #AI), with the ability to rename, merge, and create custom collections.

Concept · Tested with participants

2

Privacy control — three explicit states, one setting

"Who can see your saves": Only Me / Connections / Public, made explicit rather than left ambiguous, and placed above the saves list rather than buried in settings.

Concept · Tested with participants

3

Reminders + deadline badges

A contextual reminder concept for saved items with a time-sensitive component, like job postings — including a deadline badge and a one-tap "Apply Now" affordance.

Concept

4

Mobile-adapted layout

Horizontal folder scroll replaces the grid, filter chips replace the sidebar, and items stack in a single column — designed for mobile, not yet tested for full parity.

Concept

Phase 03 · Validation

Testing the concept with real participants

I ran usability tests with 22 participants — mid-career professionals and recruiters, recruited via LinkedIn outreach, tested asynchronously with video follow-ups.

Task	Completed	Result
Save a job post	22 of 22	100% completion
Organize saved items into folders	20 of 22	91% completion
Retrieve saved content later	19 of 22	86% completion
Comparison: default LinkedIn save vs. redesigned save	SUS survey	Redesign scored 15 points higher

93%

Overall task completion

78

SUS score

+15 pts

vs. default experience

22

Real test participants

Key findings: Folder organization was highly valued — participants repeatedly asked why it didn't already work this way. The retrieval flow noticeably reduced frustration in testing. Several participants asked for cross-device sync, a clear direction for a future iteration.

Phase 04 · Plan, Not a Result

How I'd approach implementation

As a self-directed concept, this never reached a development phase — no engineering team, no codebase. But I designed it with a real build in mind, and here's specifically how I'd approach implementation if it did move forward.

Deliverable	Format	Why This Approach
Component specs	Figma Dev Mode + annotations	Smart Collections and the privacy control are the two most complex components — precise specs reduce rebuild cycles
Cross-device sync architecture	Flagged as a phase-2 requirement	Participants asked for this unprompted in testing — it's not a nice-to-have, it's a validated request
Mobile parity pass	Full interaction audit against desktop	The mobile layout was designed but not yet tested for full parity — I'd close that gap before shipping
Privacy control default	Ships as "Only Me" by default	Explicit, safe defaults reduce the ambiguity that made the original feature hard to trust

Why this matters to me: The cross-device sync request wasn't something I designed for — it came directly from participants during testing. A real implementation plan has to account for what testing surfaced, not just what was in the original brief.

Phase 05 · Plan, Not a Result

How I'd approach launch and measurement

This was never deployed by LinkedIn — there is no real launch data, and I won't invent any. What I can share honestly is exactly what I'd want to measure and why, based on what the usability testing already told me.

What I'd Track	Why
7-day revisit rate on saved items	The core hypothesis was that better organization drives return visits — this is the number that would prove or disprove it
Folder creation rate in week 1	Testing showed 91% could organize into folders when prompted — I'd want to see adoption without a moderator present
Retrieval task success in the wild	86% completed retrieval in testing — the lowest of the three tasks, and the one most worth watching post-launch
Cross-device usage patterns	To size the phase-2 sync feature properly before committing engineering time to it

Honest end point: A tested, validated design concept — ready for exactly this kind of measurement plan, never actually deployed by LinkedIn. I'd rather show you how I think about launch than invent a launch that didn't happen.

What I Learned

Lessons from this project

User Feedback

From real usability testing sessions