There is a conversion funnel to actual paying customer. For this kind of test they are testing conversion from one step to the next. At other steps in funnels that I've seen, I've found that the increase in conversion at that step goes fairly straightforwardly to conversion to paying customer. However for email in particular at a couple of different places I've seen that improvements in clickthrough rates do not convert linearly to improvements in revenue.
While 100% more people may have clicked through, I'd be willing to bet that actual purchases went up by substantially less than 100%. This is not to discount the value of, say, a 20% increase in sales. But you really should measure that instead of clicks. (That said, people like this company like to measure clicks because it is very easy to do so.)