How A/B testing helped a price comparison app increase revenues by 28%

appiterate.com

1 pointmayankkumar2 comments

So one of the tests' result was that the control version returned more hits. Do you think too many variations can spoil the broth?

I would tend to look at it this way:

Because the control version was part of an experiment, the result is statistical proof that the control version is indeed better than the variation.

It kind of reaffirms your belief that you had when you developed the app. This reaffirmation happened only because you put it to test with another variation!!

Reply on news.ycombinator.com