Nice. I'm impressed with your pages telling people how long to run their experiments for as well. I think early stopping can be a problem, but there's a difficu

Nice. I'm impressed with your pages telling people how long to run their experiments for as well. I think early stopping can be a problem, but there's a difficult gap there. Good experiments are difficult to set up, and may be too much for a site owner who just wants to know if he should make the damn button green or red. The more you force someone to set up a complex trial, the more they'll want to go to your competitors who give them a nice flashy message saying "GREEN BUTTON WINS BY 1.4%" without all the hassle. It looks like you've got a good balance there :)