- WHO: Digital, CRO & product teams
- INDUSTRY: Airlines
- IMPACTED PAGES: Airport kiosk, Experience page, Registration page
Three experiments that show how small improvements at the moment of decision — at a kiosk, at a booking screen, on a form change whether passengers engage or move on.
Experiment 1: Redesigning the seat upgrade module at airport check-in kiosks
Hypothesis: Improving the visual design and on-screen placement of the seat upgrade recommendation module at airport check-in kiosks will increase upgrade conversion rate and reduce overall transaction time.
Results:
- +18% Increase in upgrade conversions
- -11.3% Reduction in transaction time
Takeaways:
- Featuring a carousel of premium-seat photos (vs. a plain-text list) drove 18% higher conversion rates for upgrade offers.
- By simplifying the messaging and adding clear visual options, we sped up decision-making — reducing overall transaction time by 11.3%.
- Despite the relative lift, only about 6% of all check-ins interacted with the upgrade module.
Experiment 2: Making the benefits of creating a profile clearer before booking
Hypothesis: By making the benefits of creating a profile before booking a flight clearer, you will increase the number of people that book flights while logged in.
Results:
- +6% Increase in conversion rate
Takeaways:
- Clear explanations of the benefits of creating a profile before booking increased the number of people creating a profile.
Experiment 3: Placement of profile benefits on the registration form
Hypothesis: Where you add the placement of the benefits of creating a profile will impact the conversion rate.
Results:
- +32% Increase in conversion rate
Takeaways:
- Adding the benefits of creating a profile at the end of the form increased the number of visitors creating a profile.
A carousel of seats sold more than a list of seats ever did
At a kiosk, on a booking form, in the seconds before someone decides — the details that seem small are not small. The Experimentation Ideation Agent reads your experience pages and surfaces what to test next, ranked by potential impact, so the next improvement is not a guess.