This website works best with JavaScript enabled. Learn how to enable JavaScript.

VWO Blog

A Blog Around Everything Conversion Optimization And Beyond

Never miss an update from VWO

A value for this field is required.

How a Dutch Travel Agency Increased Bookings by 33% with A/B Testing

This case study is written by Thijs van der Wal, Conversion Rate Optimization Consultant at Traffic4u, an online marketing agency that uses Visual Website Optimizer to optimize websites for their clients. This post explains how they increased bookings by 33% for Djoser, a major Dutch travel agency, and won the prestigious Gold Award at WhichTestWon’s Online Testing Awards 2014.

The test: “Take an Option”

This test was implemented on all travel details pages of and focused on travel data/price table. Within this table, visitors continued to a booking form in order to book a specific trip. In the original version (Control), a text link with a call to action button ‘Book now’ was placed next to the specific travel date.

The variation consisted of an extra link called ‘Take an option’ that opened the same booking form and captured the same information, BUT, it didn’t immediately book the visitor’s tickets. Instead, it reserved a spot in the tour for the visitor and then they had the option to cancel within 72 hours without incurring any charges. Once the 72 hours were up, the reserved spot would automatically convert into a confirmed booking.

The variation treatment with “Take an option” call-to-actions

Djoser variation with "Take an option" links

The test was chosen because Traffic4u, on the basis of data-analysis and experience, concluded that making an actual booking was too far a step for visitors of Djoser. This option was automatically converted into a booking after 72 hours which was of course clearly communicated via the booking form and the confirmation e-mail. It is possible to cancel the option within 72 hours without receiving a fine, either by phone or e-mail.


A statistically significant result (99 %) revealed that the variant resulted in a 33.1% increase in bookings (including options) during the testing period of seven weeks. Obviously, data from the backend was taken into account. Extra online bookings are a good thing, however, if these bookings are cancelled by phone or via the website, this won’t increase revenue. In this case, the (limited) amount of cancellations did not influence the significant result.

Why Does Adding ‘Take Option’ Result in Extra Bookings?

Adding ‘take option’ is a success because of its relationship with several psychological principles. An ‘option’ sounds less definite compared to a booking. With an option it is still possible to sleep on it, while your place on the tour has been reserved. The chance that an option will be cancelled is very small, because the visitor has already put in effort when entering personal information. Hence, the visitor (unconsciously) had already entered into a relationship with Djoser.


This test shows that implementing relatively small changes on a website can lead to significant improvements in conversions. However, not all tests will lead to a positive result. As an advertiser, you can come up with hundreds of tests, but just a few will lead to significant conversion improvements. On one hand, this does not matter as you can also learn from unsuccessful tests but conversely, you also want the effort invested in conversion optimization projects to actually result in conversion optimization and eventually generate more revenue. Hence, the advice for advertisers is to design a continuous process, based on the following steps:

  • Continually evaluate the effectiveness of your website (with e.g. web statistics but also user feedback)
  • Defining A/B tests on the basis of impact/effort-estimations
  • Build and conduct tests
  • Document test results and implement winning test variations

Note: WhichTestWon is a platform for A/B and multivariate testing cases. Each year WhichTestWon awards the best tests in the industry. This year, an independent jury examined 160 test cases from all over the world in different categories and selected 30 finalists. You can see the rest of the results here.

Comments (2)

Leave a Comment
  1. “The test was chosen because Traffic4u, on the basis of data-analysis and experience, concluded that making an actual booking was too far a step for visitors of Djoser.” Ah, great call. Which data indicated that users weren’t ready to actually book but might be more open to making a (sort-of-nonbinding) reservation?

    Suggested follow-up tests:

    1) This might just be a translation issue, but “Take option” sounds really weird in English. If “Optie Nemen” sounds equally strange in context, perhaps a phrase closer to “Reserve your seat” or “Save your seat” might work better. Is “Optie Nemen” a phrase that people in this language would naturally think of when they wanted to reserve a seat? (Because English-speakers would definitely NOT think of “taking an option” when they want to reserve a seat).

    2) If you find that the large majority of your users are booking through the non-binding reservations rather than through the “Boek Nu” (Book Now) links, you could A/B test placing the non-binding reservation link to the left of the Book Now link. Dutch is read from left to right, so whichever link is on the left will probably be more visible. I’ve found that showing more popular links in more visible locations tends to improve conversion rates.

    2.1) You could also try removing the “Book Now” link altogether, but I would only recommend taking this step if the overwhelming majority of users are booking through the reservation rather than through “Book Now.” I’ve found that removing rarely-used links tends to improve conversion rates, but overall I suspect that test

    removing the “Book Now” links or placing the “Optie Nemen” / optional reservation links to the left of the “Boek Nu” / Book Now links so that most visitors will see the optional reservation option before the Book Now option. (Dutch is read from left to right, so whichever link is on the left will probably be more visible). I’ve found that removing rarely-used links or making them less prominent than more-commonly-used links tends to improve conversion rates.

  2. Hi Brian,

    Thanks for your comment in which you provide some good suggestions en asked some good questions, we appreciate your feedback.

    Perhaps the following test details will give you a clearer view of the test. On Djoser it was in the control situation also possible to ‘take an option’. BUT, this was only possible AFTER you’ve completed the full booking form. On the last page of the booking process there was a tiny checkbox to get an option of the booking instead of a real booking. Besides that, we saw some competitors offering this same service.

    About the translation of “Optie nemen” I think your right: “Take an option” or “Get option” might not be the best translation. Perhaps “Reserve your seat” catches the meaning better (don’t forget to test that). Placing the “Reserve your seat” link on the left is a nice one to test, because of the priming and contrast principles.

    Removing a button can also be helpful (Nielsen’s: minimalistic design). But in this specific case we focused on the group of people who are not willing to make a definitive booking yet, but can be persuaded to make a nonbinding reservation.

    Conclusion: don’t make assumptions, always act on data and always test your hypotheses.

    Kind regards,


Comments are closed.

Contact Us / Login