Mobile appInspiration in the pocket
Discover destinations, save ideas and return when a trip becomes real.
LUXTJ · Luxury travel website
From messy ideas to a product people could understand.
The first version looked complete, but people struggled to understand key features and finish important journeys. I used that failure to redesign the product around learnability, decision speed and confidence.
- Role
- Lead Product & UX Designer
- Ownership
- Research, flows, UI and validation
- Timeline
- 12 months
- Team
- Founder, stakeholders and engineering
- Scope
- Website + connected product system
- Read time
- 9 minutes
2 min → 30–50 secTravel Calendar comprehension
~60%Checkout drop-off surfaced
22% fasterCheckout completion
+70%Social-profile engagement
At a glance90-second summary
- The problem
The first version asked people to configure a Personal Travel Calendar, interpret dense checkout and navigate an unstructured profile before showing value.
- What I changed
I rebuilt four journeys around one clear starting point, progressive disclosure and the next decision.
- What resulted
People understood the Personal Travel Calendar in 30–50 sec, checkout completed 22% faster and profile engagement rose 70%.
One travel idea. Several moments that had to feel like one journey.
Mobile app
Customer websiteSpace to compare and plan
Explore, evaluate and assemble a complex trip with more context.
Partner extranetOperations behind the promise
Manage content, availability, pricing and bookings behind the customer experience.
The uncomfortable start
My first mistake was designing for completeness before comprehension.
The initial product tried to expose the richness of luxury travel immediately. In practice, that richness became work. New concepts looked like forms, checkout became an information wall, and profile options appeared without a reason to begin.
What changed my directionA finished-looking screen can still fail if a person cannot explain what it is for, what to do first or what happens next.
If all the useful options were visible, people would feel informed and in control.
More options increased interpretation time, backtracking and uncertainty.
Move from screen completeness to progressive understanding and decision support.
Every screen had to answer the next important question.
What was happeningQuestion left openDesign response
Disconnected search screensWhat have I already chosen?Persistent trip context
Dense result cardsWhat makes this right for me?Progressive hierarchy
Totals appeared lateWhat will this decision cost?Running price and consequences
Errors at submissionHow do I recover?Inline validation and editable review
How I worked through the mess
Failure became a repeatable design loop.
I stopped treating each weak screen as an isolated UI issue. The same loop helped me find the underlying behaviour, redesign the journey and learn from the result.
- 01
Observe
Watch where understanding, confidence or momentum breaks.
- 02
Name the risk
Turn visible friction into one testable design question.
- 03
Study patterns
Benchmark comparable journeys and the behaviour behind them.
- 04
Simplify the path
Reorder information around the user’s next decision.
- 05
Learn again
Compare behaviour after redesign and carry the lesson forward.
What came out of the loop
Four rules kept the product coherent as it grew.
Explain before asking
Introduce the value of a new feature before asking people to complete a form.
Keep context close
Keep dates, travellers, selections and price visible while decisions are being made.
Prioritise the next decision
Put the evidence needed now first; keep deeper detail available progressively.
Design the recovery
Make edits, missing information and changed plans understandable and recoverable.
Four contributions
What failed, what I changed and what I learned.
Each chapter starts with the friction, not the final mock-up, so the design decisions and my contribution stay visible.
A calendar of dates did not yet feel like a reason to travel.
How might personal milestones become useful travel inspiration without making setup feel like administration?
The original Personal Travel Calendar opened with an unfamiliar setup flow. Before adding anything, people spent about two minutes working out what the feature was for, which dates belonged there and what would happen next.
I studied progressive disclosure, demonstration-led onboarding and familiar calendar patterns. The core issue was not entering a date, it was asking people to organise information before showing how the product could turn a birthday, anniversary or meaningful moment into a travel plan.
I introduced the feature through recognisable personal moments, showed the travel-planning benefit first and created one clear path for adding a date. Recommendations and next steps then appeared progressively, keeping the calendar useful rather than making it another form to maintain.
30–50 secto understand the Personal Travel Calendar, down from about 2 minutes
What I learnedShow how a personal date becomes a meaningful plan before asking people to maintain it.
More information created less confidence.
How do you keep essential trip details visible without making checkout feel endless?
The original checkout brought every decision, policy and trip detail into one long page. Users had to scan through too much information at once, contributing to a drop-off rate of nearly 60%.
I audited checkout patterns across travel platforms, comparing how they sequence travellers, itinerary details, pricing, policies and payment. The strongest examples reduced memory load without hiding consequences.
I reorganised checkout into decision-sized groups, kept the trip summary and price visible, made sections editable and moved supporting detail closer to the moment it was needed.
22%reduction in checkout completion time after redesign
What I learnedFast checkout is created by hierarchy and recovery, not by removing every detail.
The details were present, but not decision-ready.
What does a traveller need to know before a hotel feels safe to choose?
Hotel selection took the longest part of the journey. Important differences were difficult to scan, forcing people to open, compare and revisit options repeatedly.
I reviewed hotel product-detail pages across travel websites and mapped the evidence people use to decide: location, inclusions, room differences, cancellation, price and the path forward.
I redesigned the description and room-selection pages around comparison. High-value evidence appears first; richer detail remains available progressively; price, policies and actions stay connected.
Shorter pathfrom hotel consideration to checkout, observed after redesign
What I learnedA useful detail page does not show everything equally, it answers the next decision first.
A list of features is not a reason to start.
How can profile creation feel like progress rather than administration?
Profile interest was close to zero. Multiple pages and features were presented at once, with no obvious first step and no visible reason to complete them.
I studied the existing behaviour, mapped where people stopped and explored the trigger–action–reward ideas in Hooked. The opportunity was to create momentum, not add another reminder.
I introduced one clear starting point, showed visible progress and used a trigger–action–reward structure to surface genuine travel-personalisation value, not manufacture urgency.
+70%time spent in the social profile experience
What I learnedEngagement grows when every small action reveals a meaningful next benefit.
My contribution
I carried the work from ambiguity into delivery.
I clarified the product idea, studied user behaviour, mapped journeys, explored patterns, redesigned the information hierarchy and worked with developers to resolve edge cases before implementation.
Converted loosely defined travel ideas into clear user problems and measurable questions.
Combined observed friction, comparative analysis and behavioural patterns into design direction.
Redesigned core flows, page hierarchy, component behaviour, validation and recovery states.
Maintained a shared library across three platforms and worked closely with engineering through handoff.
The smaller states carried real product risk too.
01Flight selection
Shared patterns make duration, stops and fare conditions easier to compare.
02Hotel discovery
Filters, map context and richer results support a confident shortlist.
03Room selection
Policies, inclusions, price and the next action stay connected.
04Trip planning entry
A clear introduction explains the value before asking people to add information.
Making the learning reusable
The redesign became a shared product language.
The same principles, including progressive disclosure, persistent context, visible consequences and clear recovery, were carried into reusable patterns across customer web, mobile and the partner extranet.
Reflection
The best outcome was a better way of designing.
Comprehension comes before conversion.
If people cannot explain the value, optimising the action is premature.
Hierarchy is a product decision.
What appears first shapes confidence, speed and the quality of every choice after it.
Failure needs a diagnosis, not decoration.
Changing the visual layer would not have fixed unclear concepts or missing starting points.
Instrument the next version early.
I now define the behaviour to measure alongside the experience, not after the design is finished.
Evidence note. Figures shown here reflect project observations and product data available during the work. Where exact instrumentation was limited, I describe the result directionally rather than inventing precision.
Want to talk through the messy middle?
I can walk through the failed directions, research choices, alternatives and delivery decisions in a portfolio conversation.