Case study · Invia Flights · 2025 to 2026
Three features for an online travel agency, each answering a customer's doubt at the moment it appears. All A/B tested, all measured against conversion.
Context
Invia Flights is one of the largest online travel agencies in the German market, with brands like fluege.de and airline-direct.de on web, iOS and Android.
I owned three features at three points of the funnel. Each one answers a question that used to cost us bookings.
1 · Search results
"Is this the best offer, and is the member price worth it?"
Problem. The old card had one big button that selected the member price. The normal price hid in a small text link. Customer reviews on Trustpilot and drop-offs in the booking funnel, analyzed in session recordings and analytics, showed the cost: customers felt pushed.
Tension. Making the normal price visible could cost member sign-ups, and the Product Owner was concerned about that. So before designing, we agreed on bookings as the success metric and member clicks as the guardrail.
Hypothesis. If both prices carry equal weight, customers feel less pushed and book more, even if fewer pick the member price.
1 Earlier
One button, one path. "Angebot wählen" picks the member price. The normal price is a small link.
2 Before my redesign
Same push, more height. The big button still picks the member price, and the card grew taller.
3 My redesign · live today
Two equal choices. Both prices are their own button, in almost a third less height than version 2.
I turned the card into an honest choice: two equal price buttons, side by side. I defined UX KPIs for the layout, A/B tested it against the previous card and checked real behavior in session recordings. I built the box as a modular component, so new features and layouts can be tested slot by slot without a redesign. The chip slot was planned from the start for features still to come, like the Flex Ticket in the next chapter.
One box, four slots: top, itinerary, price actions and footer. Each slot changes on its own, so new features were added without a redesign and button layouts could be A/B tested. The tour plays one round, then it's your turn.
Try it: hover a stop, click a price
Result
A measurable conversion uplift, despite fewer upsell clicks. A less pushy offer sold more flights, and it is still live today. The test also surfaced ideas for the next round, ready to be tested in their own slots.
2 · Booking
"What if my plans change?"
A flexible ticket lets customers cancel shortly before departure and get most of the price back. We A/B tested two layouts, iterated on the winner, then adapted it for the apps. Thanks to the chip slot in the Offer Box, the Flex Ticket also shows up right in the search results, without touching the card's design.
Variant A · Vertical
Stacked like the other products on the page. The safe default, and our control.
Variant B · Horizontal Won
Ran in parallel to A. Our hypothesis: side by side, customers compare the benefits faster and pick the ticket that fits them. With the benefits lined up row by row, customers decided significantly better.
Iteration · B with hint
Once someone picks the Standard ticket, a short hint reminds them that plans can change. It helps people judge whether the flexible ticket fits them, and it raised the take rate.
Branch
We redesigned the booking funnel in the iOS and Android apps. Variant A's vertical structure became the bridge to the web: same content, same order, built within Apple's Human Interface Guidelines and Material Design.
iOS
Android
Result
About 1 in 3 customers chose the Flex Ticket, and conversion rose by a high single-digit percentage. People booked who would otherwise have waited.
We also aligned design system decisions with the mobile development team. Web and apps now feel more consistent, and both teams avoid duplicate work.
3 · Payment
"What if the price drops after I book?"
If the fare drops after booking, customers get part of the saving back as a voucher. We tested where in the funnel the offer belongs, then made the mechanics easy to understand.
Variant A · Early in the funnel
Right after the contact details, next to offers like the Flex Ticket. The assumption: early means more attention.
Variant B · At the payment step Won
Our counter-hypothesis: once the other add-ons are decided, the offer gets full attention at payment. So we moved it there. With fewer choices on screen, customers decided with less cognitive load, and more of them took it.
UX writing · How it works
The link under the box opens a short dialog: three numbered steps first, the fine print after. Customers see what they get before they decide.
Result
Over a third of customers booked with the Price Guarantee.
What I would change next
Today the box ships pre-selected. I would make it an active choice: a guarantee customers pick themselves builds more trust, and trust is what all three features are about.
What I took away
Curious about the tests, the iterations and what didn't work? That's the part I like to tell in person.
Let's talkNext case study
