It also hinges heavily on how you're tracking these A/B results. There's less certainty and value in simply saying "Group A ended up booking more" —  if metrics

It also hinges heavily on how you're tracking these A/B results. There's less certainty and value in simply saying "Group A ended up booking more" —  if metrics indicated a statistically higher percent of users complete the purchase at that particular page with the single change on it. Now that we know it worked for Kayak and not another company, we can delve into why to further inform everyone's future decisions. IMO a padlock with more pure language would do better, but only another A/B test by Kayak or similar company would validate this.