What’s an A/B test you’ve performed that surprised you and why?

An A/B test I’ve performed in my previous job was when we had to send out different versions of announcements of a webinar my former company was hosting. We tested it first with the Marketing and Creative team using different colors for the email layout. Once we established that version B performed higher than version A, we went ahead and used version B. Surprsingly, our hypothesis was correct. Version B perfomed really well and we got a high engagement rate and click-through rate in our email.