Hi guys Keen to use auto-allocate to test a new feature on one of our critical metrics however docs suggest Manual allocation is better for 'characterizing performance' but I cant understand why can someone elaborate?
@richardlee the easiest way to describe it is do you just care that B is better than A or do you care precisely how much better B is than A as well as the other experience variations? Auto-Allocate will find the winner and start shifting traffic over faster.