Only as a side note, and of course IMHO, A/B testing may not necessarily - even if sometimes it happens - end up with actually improving user experience. What

Only as a side note, and of course IMHO, A/B testing may not necessarily - even if sometimes it happens - end up with actually improving user experience. What it does - once a result is chosen - is to "shift" user experience towards what an unqualified numerical majority likes more (which not always is the same as better), and while the experiment is running it risks to irk a lot of users. To the scope of making the whatever more popular. it is of course appropriate, but it shouldn't be called "better" or "improvement" unless it is actually better or it really is an improvement. About the specific on dark patterns, you must I believe, re-consider your (or your firm's) approach, as you say "luckily " it didn't work this time, but that doesn't mean that the next dark pattern will not work, possibly surprisingly well. Dark patterns exist exactly because they are successful (in sheer numbers of subscriptions, conversions, etc.), even if often only on the short or very short term. We cannot deny this. Many companies use dark patterns and are successful until they cross an invisible line that - all of a sudden - makes their customers angry at them. In a perfect world customers at the very first fraction of a hint that maybe a dark pattern is used should get angry and leave. But this is not what happens in real life, they come and stay until they don't. So you won't be able to convince anyone that dark patterns are "bad" (as in the short term they do work) if not by scaring them about the possible long term effects or by making ethical considerations (historically things like a company motto "Don't do evil" didn't work all that well)