> I wouldn't expect much change if there are no numbers that tell "plain converts better". The null hypothesis should be that plain and rich convert no differe

> I wouldn't expect much change if there are no numbers that tell "plain converts better". The null hypothesis should be that plain and rich convert no different. I'd send plain in the absence of any convincing evidence to send rich.