ParkShare
The Airbnb for Parking. A peer-to-peer marketplace that turns unused driveways and private lots into bookable parking, starting where the crisis hits hardest: university campuses, dense urban cores, and event venues.
For the student circling a campus lot at 8:52 AM, parking is a daily lottery. Research showed people don't need more options. They need certainty, safety, and control. I led research and product design across both sides of the marketplace.
Product Snapshot
- • Problem: Parking is a daily lottery that eats 10–30 minutes, driven by uncertainty more than scarcity.
- • Solution: A peer-to-peer marketplace that turns unused driveways into bookable, trust-verified spots, designed for predictability instead of search.
- • Research: 6 in-depth interviews, 5 personas, 4 usability participants.
- • Result: SEQ 6.23/7 with 100% task completion; Color UI significantly improved rebooking (p≈.031).
- • Stage: Hi-Fi and comparative study complete, pilot planning.
- • Limits: Exploratory sample (n=4) with a Midwest skew, documented rather than hidden.
In progress: next up is a pilot launch.
- Role Product Designer, UX Researcher, Prompt Engineer
- Timeline Oct 2025 – Present
- Methods Contextual Inquiry, Thematic Analysis, Usability Testing, Comparative A/B Study, AI-Augmented Coding
- Tools Figma, GPT-5, Notion, Miro
- Status Hi-Fi Complete • Comparative Study Complete • Pilot Planning
6.23/7
Overall SEQ Score
100%
Task Completion
4
Usability Testers
6
In-Depth Interviews
5
Research Personas
5
Tasks Tested
The hard calls, and how I made them
Every flow below starts from a research theme, not a feature idea. People arrived two hours early just to secure a space: the problem was uncertainty, not search.
"Will there be parking?" was an open question. Availability uncertainty dominated interviews, and nothing answered it before the drive.
How I solved itA predictive availability map with confidence scores turns that open question into an answerable one.
Principle: Design for certainty, not for search.Peer-to-peer parking asks for trust strangers haven't earned, and comparing options carried a real cognitive cost.
How I solved itResults are colour-coded by trust level; spot cards show proximity, price, safety score and legal status, the four criteria research identified.
Principle: Make the comparison cheap and the risk visible.Arrival was the anxiety peak. Journey maps put the worst moment on arrival, not booking: "Am I in the right place?"
How I solved itAn arrival-zone notification at roughly 50m delivers access details and host confirmation at that exact moment.
Principle: Intervene where the journey map shows the spike.Commuters were re-deciding the same thing daily, and the repetition wore them down.
How I solved itOne-tap Book Again and Auto-Reserve Weekdays turn parking from a daily decision into background automation.
Principle: Remove the decision, not just the friction.Design principles & psychology at work
Research insights encoded directly into the interface:
- Persistent contextual anchoringDriver mode (blue) and Host mode (purple) are color-segmented to answer "which mode am I in?" peripherally, fixing mode confusion seen in early testing.
- Borrowed mental modelsNavigation leans on familiar maps-style patterns to minimize learning curve, the flow scored 7/7 SEQ.
- Progressive trust-buildingThe host flow leads with income visibility and ends with projected earnings, because hosts need the "why" before the "how".
- Designed for stressed hands44px minimum tap targets reflect mobile use when time is short and stakes are real.
01The Challenge: A Daily Lottery Nobody Wants to Play
"If you get a parking spot, it's like lottery."
- Participant
"As a girl, I feel safer when the lot is more full and there's more light."
- Participant
Picture a student circling a campus lot at 8:47 AM with a 9:00 class and a permit that sold out months ago. Event-goers face worse: moving cars before midnight to avoid towing, or walking 10+ minutes through sub-zero temperatures. Parking has become a tax on everything it's supposed to enable.
The frustration isn't scarcity; it's information asymmetry. Rules shift by day and event, and people who felt "in the clear" still feared boots and tickets.
ParkShare's Systematic Response
The overlooked insight: cities are full of private parking that sits empty while drivers circle. ParkShare turns that inventory into a trusted, bookable network, with less circling, lower CO₂, and wider access to safe parking. The real challenge was cognitive: shifting parking from "lottery" to "reservation."
02Research Approach: Methods and Protocols
Research Goals
- • How drivers search, select, and pay, and what they trade off under pressure
- • The pain points, safety concerns, and rule confusion current tools miss
- • Whether people would trust a peer-to-peer driveway model, and on what conditions
- • Core flows, tested with scenario-based tasks
- • Findings translated into strategy, HMWs, and prioritized opportunities
Participants
6 interview participants: students, commuters, and urban drivers across Ames, Des Moines, Chicago, and Massachusetts, aged early 20s to mid-30s.
4 usability participants for the mid-fidelity prototype, all comfortable with navigation and parking apps like ParkMobile.
Methods Applied
Contextual and semi-structured interviews • Codebook development and applied coding • Thematic analysis (6 key themes) • Insight statements with impact/confidence rankings • HMW generation and opportunity–solution tree • 5 research-backed personas • Lo-Fi usability testing with SEQ ratings • Hi-Fi comparative study (within-subjects) with UEQ • Bias and rigor audit
Research Protocols
📝 Discovery Interview Protocol
- Duration: 20 - 25 minutes
- Format: Semi-structured, remote
- Recording: Audio w/ consent
- Structure:
- - Warm-up & context setting
- - Parking behavior exploration
- - Pain point deep-dive
- - P2P concept reaction
- - Wrap-up & debrief
🔲 Lo-Fi Testing Protocol
- Duration: 20 - 25 minutes
- Format: Think-aloud, moderated
- Tasks: 5 scenario-based
- Metrics:
- - Task completion (Y/N)
- - SEQ per task (1 - 7)
- - Hesitation patterns
- - Label comprehension
- - Flow recoverability
🎨 Hi-Fi Comparative Protocol
- Duration: 25–30 minutes
- Design: Within-subjects, counterbalanced
- Variants: Greyscale vs Color
- Data Collected:
- - Completion time (seconds)
- - SEQ difficulty (1–7)
- - UEQ experience scoring
- - Think-aloud insights
- - Variant preference rationale
03Early Sketches & Ideation
Before digital wireframes, I sketched the system on paper: mapping end-to-end flows, stress-testing hierarchy, and finding where users hesitate or lose confidence.
Search & Book Flow Exploration
Figure: Discovery to payment, testing proximity-first scanning.
Navigation & Arrival Flow
Figure: Arrival confirmation beyond GPS arrival.
Rebooking & Smart Prefill Concepts
Figure: Recognition-based rebooking.
Host Onboarding & Listing Flow
Figure: Earnings-first host onboarding.
04Key Findings: Six Themes from the Field
1. Reliable Spots Beat the Lottery
Participants arrived 1–2 hours early to secure a space, trading time for certainty. Implication: guarantees and confidence indicators.
2. Users Pay for Proximity & Time
"If it's near to that event, I'm willing to pay for it." A rational time-value trade. Implication: frame cost as time saved.
3. Trust Rises with Safety Signals
Lighting, occupancy, and host verification shape perceived safety; P2P has to feel like a vetted network. Implication: safety scores, verified badges, nighttime photos.
4. Rule Clarity Reduces Anxiety
Ambiguous signs keep people anxious even when they've done everything right. Implication: legality overlays and towing countdowns.
5. Frictionless Flow = Value
Users avoid lots that are "more fuss than it's worth." Implication: skill-based filters and fast-exit routing.
6. Emotional Fatigue Is Real
"Why should this be a hassle for me?" Fatigue compounds. Implication: every interaction should remove load.
05Five Research-Backed Personas
Five personas emerged from interview data, each grounded in direct quotes.
Lena
Time-Pressed Commuter
Minimize hunting. Treats parking as a scheduling problem, not a search problem.
Opportunity: Predictive map, "Guaranteed Spot" reservations, auto-reserve weekdays.
Caleb
Budget-Conscious Student
Avoid fees. Trades walking distance for money saved.
Opportunity: Transparent pricing, legality indicators, student discounts.
Maya
Cautious Night Driver
Stay safe at night. Filters every option through safety first.
Opportunity: Safety score, "Host Verified" badges, nighttime photos.
Ryan
Weekend Event-Goer
Park close, exit fast. Pays a premium only when rules are clear.
Opportunity: Event-specific alerts, "Fast Exit" routing, towing timers.
Denise
Reluctant Host Homeowner
Earn passively, with control. Hesitant until liability and legitimacy are clear.
Opportunity: Host Protection dashboard, verified renter identities, earnings projections.
Research Persona Framework
Figure: The five personas, derived from contextual inquiry and thematic analysis.
06Empathy and Journey Maps: Where Anxiety Peaks
Empathy maps captured what each primary persona says, thinks, does, and feels. Journey maps then showed that emotion doesn't track touchpoints: anxiety peaks at search and arrival, relief at confirmation and departure. ParkShare's interventions target the anxious stretches.
Driver Empathy Map (Lena)
Figure: Driver empathy map. The relief when plans work is what ParkShare has to deliver.
Host Empathy Map (Denise)
Figure: Host empathy map. Each concern gets evidence and structure, not reassurance.
Journey Maps
Driver Journey Map: Search → Reserve → Navigate → Park → Depart
Figure: The confidence meter and arrival confirmation intercept the two anxiety peaks.
Host Journey Map: Consider → List → Manage → Earn → Expand
Figure: Insurance clarity and earnings projections are the host's conversion points.
07User Flows: Mapping Every Decision Point
User flows exposed friction that isolated screens hid. Each decision point was mapped against a research theme; the four flows below became the hard calls described at the top.
Driver Journey: Search → Compare → Reserve → Navigate
Figure: Proximity and price surface early, because research showed they drive selection under time pressure.
Navigating to Spot Flow
Figure: Navigation, arrival detection, and parking confirmation. Scored 7/7 SEQ by borrowing familiar maps patterns.
Host Journey: Onboard → List → Manage → Earn
Figure: Earnings projection at step 4, because hosts need to see potential income before committing effort.
Rebook & Favorites Flow
Figure: Favorites and one-tap rebooking. This flow later showed a significant improvement with color UI (p≈.031).
08Design System: Encoding Mode and Trust
Visual Direction & Mood
The mood board set the tone before any tokens: calm, trustworthy, and dependable, for a task that is inherently stressful.
Mood Board: Trust, Control, and Calm
Figure: Calm blues signal reliability for drivers; grounded purples signal ownership for hosts.
Foundations: Color, Type Scale & Iconography
Figure: Role-based palettes, type scale, and iconography, so mode, urgency, and state read at a glance.
Design Elements & Interaction Patterns
Figure: Reusable patterns for maps, result and listing cards, completion states, and primary CTAs.
09Lo-Fi Prototype and Testing: Structure Before Polish
The mid-fidelity Figma prototype tested structure before polish. 4 participants reached SEQ 6.23/7 with 100% task completion, and their feedback showed where my intent and their interpretation diverged.
Onboarding Flow: Welcome → Account Setup → Preferences → Ready to Park
Figure: Onboarding. Users wanted value before registering.
Flow 1: Find & Reserve a Guaranteed Spot (Driver Journey)
Figure: Find & Reserve, SEQ 5/7.
Flow 2: Navigate & Park (Directions → Arrival → Confirmation)
Figure: Navigate & Park, SEQ 7/7.
Flow 3: Host Lists a New Space (Onboard → List → Publish)
Figure: Host listing, SEQ 7/7.
Flow 4: Rebook & Smart Scheduling (Favorites → Auto-Reserve)
Figure: Rebook, SEQ 4/7, the lowest, which shaped the Hi-Fi work.
Lo-Fi Usability Testing Results
4 participants, 5 core tasks: reserve a spot, navigate and confirm arrival, list a space, rebook a favorite, and set up weekday auto-reserve. Each was followed by an SEQ rating and open reflection.
Task-Level SEQ Scores (Lo-Fi)
What Worked
- ✓ "The UI is simple" - P4
- ✓ "It met my expectations... similar to Google Maps and Apple Maps, no learning curve" - P4
- ✓ Navigation feature: "Navigator is going to be so useful because sometimes you forget where you're parked" - P4
- ✓ Host listing rated 7/7: "Clear prompts that told the user what next to do"
Design Changes Needed
- ✗ Search needs destination entry: "You could be going to a different place"
- ✗ Rebook flow felt incomplete: "I feel like there's a whole path that's missing"
- ✗ Auto-reserve should pre-fill the last booking instead of "having to repeat information"
- ✗ Filters for price, distance, and nearby facilities
- ✗ A booking receipt as proof of payment
- ✗ Simpler photo upload at arrival: "that's a lot of stress"
Feature Suggestions from Testing
Book for a Friend: "Can you book for someone?"
Nearby Facilities: show food and gas nearby.
Confidence Meter: valued, but the label needed work.
"It's a nice app... saves so many people the stress of having to find parking... It's a win-win situation."
- P4, Usability Testing Session
10Hi-Fi: Greyscale Baseline vs Role-Segmented Color
The Greyscale UI was the study's baseline, isolating structure (hierarchy, layout, copy) from visual treatment. The Color UI put the same architecture into a role-segmented system: blue for Driver, purple for Host, semantic accents for status and urgency.
Onboarding, Search & Booking (Greyscale)
Figure: Greyscale onboarding and first booking.
Navigation & Arrival Flow (Greyscale)
Figure: Greyscale navigation and arrival.
Host Onboarding & Listing Creation (Greyscale)
Figure: Greyscale host setup, where hosts lost track of mode.
Automated Rebooking (Greyscale Baseline)
Figure: Greyscale rebooking, where the gap was largest.
Greyscale UI Interactive Walkthrough
Video: Full walkthrough of the greyscale baseline.
Hi-Fi Prototype: Color UI (Role-Segmented)
Onboarding, Search & Booking (Color)
Figure: Color onboarding and booking, with semantic accents as quick-read cues.
Navigation & Arrival Flow (Color)
Figure: Color navigation; participants double-checked labels less.
Automated Rebooking (Color)
Figure: Color rebooking, the largest gain.
Host Onboarding & Listing Creation (Color)
Figure: Host purple removed mode confusion.
The system in motion
The same two flows, end to end: a driver books, pays, navigates, and parks, then a host lists a driveway. Greyscale baseline on the left, role-segmented color on the right.
Color UI Interactive Walkthrough
Video: Full walkthrough of the colour-segmented variant.
11Comparative Study: Where Color Made a Measurable Difference
Could color shift users from recall ("which mode am I in?") to recognition ("I can see which mode I'm in")? Four participants did identical tasks in both variants, counterbalanced, to find where, why, and for whom color matters.
Quantitative Results
Task Completion Time: Greyscale vs Color UI
Figure: Mean completion times across five tasks. Rebook shows the largest gap, and it's significant (p≈.031); Reserve is directional (p≈.087). With n=4, this is exploratory evidence, not a definitive result.
Experience & Perception Results (SEQ + UEQ)
User Experience Questionnaire (UEQ) Scores: Greyscale vs Color UI
Figure: Color UI gained most in Efficiency, Attractiveness, and Dependability; Novelty was mixed. SEQ ratings aligned.
Color UI Advantages (UEQ):
- ↑ Efficiency: less time orienting, more time executing
- ↑ Attractiveness: preferred, but secondary to function
- ↑ Dependability: clearer Driver vs Host awareness, the shift the design targeted
- ↑ Perspicuity: easier to understand at a glance
Mixed or Neutral Findings:
- ⚖ Novelty: expectations varied
- - Learning curve: low friction in both; the structure held
Qualitative Results
Color UI Preference Themes:
- • "I immediately knew where I was in the app"
- • "The colors helped me find what I needed faster"
- • "Green for hosting makes sense, like earning money"
- • "Less confusing when switching between modes"
Greyscale Observations:
- • "Clean but sometimes I forgot which tab I was on"
- • "Had to read labels more carefully"
- • "Works fine but the other version felt more intuitive"
- • "Professional look, but less memorable"
Qualitative feedback matched the numbers: color improved mode awareness through efficiency, not preference, most in repeat-use flows.
12AI Collaboration as Research Accelerator
As researcher and prompt engineer, I used GPT-5 to turn raw interviews into a full research stack. Every output was validated against source data; interpretation stayed with me.
Prompt 1: Codebook
12–20 codes with definitions and example quotes; I chose which to keep.
Prompt 2: Applied Coding
Applied across 6 transcripts, surfacing new codes like Lot_Egress_Friction and P2P_Cultural_Legitimacy_Barrier, each checked against raw segments.
Prompt 3: Themes
6 themes: Scarcity, Cost Trade-offs, Trust & Safety, Rule Anxiety, Mobility UX, Emotional Fatigue, each traced back to evidence.
Prompt 4: Insight Generation
Insight statements with impact and confidence ratings, plus How Might We questions for each theme.
Prompt 5: Persona Synthesis
5 personas, quotes verified against transcripts.
Prompt 6: Journey Mapping
Emotional journeys tied to quotes; AI drafted, I refined.
Prompt 7: Opportunity–Solution Tree
Needs → sub-opportunities → solutions, prioritized by strength of evidence.
Prompt 8: Bias & Rigor Audit
Checked for sampling bias, leading questions, overreach, and confirmatory reasoning, and documented the limits.
Validation & Rigor
Everything was cross-checked against transcripts and my session notes. The audit flagged a Midwest skew, a 20s–30s cluster, and confirmatory risk, documented rather than hidden, which is where the pilot and wider sampling come next.