My gut reaction agrees, but I have seen times, especially at 'telly' where we were able to do a quick test and prove that users were interested in a feature. H

My gut reaction agrees, but I have seen times, especially at 'telly' where we were able to do a quick test and prove that users were interested in a feature. However in that case we were both watching data and only running the test for a few hours. I'm wondering when there are best practices to do this, and maybe when not to do it.