Case Study
Redesigned Costsaver’s digital experience across the homepage, trip pages, destinations, itineraries, and visual identity to make booking easier. The result was fewer visitors, but more of them booking.
RoleSenior Product Designer
CompanyBS&T (In-house)
ScopeResearch, Prototyping, Design Systems, Production Design
Timeline2024 to 2025
Transactions
+50%
YoYMobile booking share
+3pts
YoYTrip-page engagement
+30.3%
YoYCostsaver runs affordable guided tours across the world and consistently brings in over 400,000 visits a month. Despite performing well commercially, the site had started to feel too close to other brands within the group, sharing many of the same components, layouts and interaction patterns.
At the same time, the brand team was working on updated print guidelines and a brochure redesign with another agency. My role was to lead how that new direction came to life digitally. I redesigned the visual language of the website, improved key customer journeys, and created a more flexible design foundation the team could continue building on properly over time.
ConstraintWCAG 2.1 AA compliance as the absolute minimum
ConstraintStakeholder pressure to launch quickly, with limited room for user testing and validation
ConstraintInconsistencies between approved designs and final development output
ConstraintSeparate design and development teams with indirect communication
Worked alongside a UX researcher to synthesise interviews, analytics, session recordings, and stakeholder feedback. Four clear issues kept surfacing.
My contribution
Research planning support
Helped shape the research plan and decide what to look for.
Analytics and behavioural analysis
Read analytics and session recordings for repeat patterns and friction.
UX synthesis
Grouped the signals into the four findings shown below.
Translating research into design
Turned each finding into a specific change to the journey.
Finding 01
The homepage led with a search bar, but most visitors didn’t know where they were going yet. The experience assumed intent too early.
Finding 02
Trip pages split attention across imagery, pricing, and details, making it harder to quickly understand value. Users had to scan and piece information together before making a decision.
Finding 03
Choice Experiences were a core part of the product, but there was no clear way to explore them, compare options, or understand cost without doing the maths manually.
Finding 04
Typography and colour had drifted across the site. There was no shared system in place, which made consistency hard to maintain and slowed the team down.
Mapped the end-to-end journey from discovery to booking, focusing on where the experience was breaking down and what needed to change.
Change 01 · Discovery
User problem
The homepage led with a search bar, but most arrivals hadn’t decided where to go. The search-led homepage killed the browse mood and pushed casual visitors away.
Key moves
Change 02 · Comparison
User problem
Travellers had to scan dozens of departures to find the one that matched their flexibility. Anyone shopping last-minute or on a budget had no shortcut, just a long list.
Key moves
Change 03 · Decision-making
User problem
Trip pages buried pricing and key details under long imagery galleries. Travellers had to scroll past inspiration to reach the information that actually drove their decision.
Key moves
Change 04 · Personalisation
User problem
Choice Experiences were a core part of the product but scattered through the itinerary, with no clear way to compare options or see the running cost of selections. Most travellers never realised they could personalise the trip at all.
Key moves
Interaction states
Added. Hearting saves it to the wishlist.
Replaced. Picking an alternative swaps the existing one.
Removed. Unhearting clears it from the wishlist.
Change 05
The visual language had drifted, making consistency hard to maintain. A clearer system for type, colour, and components was introduced to support accessibility and speed up delivery.
The redesign focused on intent-driven travellers rather than casual browsing. Sessions dropped about a fifth year-on-year. The visitors who did arrive converted at a much higher rate, and revenue rose 54%.
Organic traffic, Sep to Dec year-on-year
Revenue
+54.3%
Transactions
+50%
Avg. session duration
+48.9%
Organic traffic, year-on-year. Fewer users, more revenue.
All traffic, full 8 months year-on-year
Engagement improved across the core templates the redesign touched. Time on the homepage more than doubled, with the biggest jump showing up on content pages.
Time on homepage
02:22
from 01:01 +132.8%
Time on content pages
01:44
from 00:31 +235.5%
Time on trip pages
02:39
from 02:02 +30.3%
Aug 2024 to Apr 2025 (before)
Same months, a year later
Secondary signals
Mobile bookings share grew
27.1% → 30.2%, up 3 points
Most users browse on mobile but convert on desktop. The simplified flow helped close that gap and moved more bookings onto mobile.
Direct traffic pulled its weight
22.9% → 31.0%, up 8 points
Users arriving with intent were more likely to convert, suggesting the experience was doing a better job of supporting decision-making.
Caveat
Some of this period includes changes in paid channel mix and seasonal variation. The cleanest signal comes from page-level engagement, where comparisons are consistent before and after launch.
Secondary elements like dropdowns, modals, and tour details were missed first time around. Mapping the full set upfront would have caught the gaps.
Design decisions were relayed through another team before reaching developers, introducing avoidable friction and misalignment. It’s something I pushed for across projects but couldn’t shift, reinforcing how important it is to secure direct collaboration early.
Top-line metrics are noisy. Channel mix, seasonality, and paid spend move independently of design. The clearest signal came from page-level engagement on the templates we changed. I’d define those metrics before launch next time.
The brief covered a lot of ground. Focusing more tightly on the highest-impact journeys upfront would have made it easier to go deeper where it mattered most.
I partner with product teams who want UX decisions you can defend in the data.