Quote:
Originally Posted by studiocritic
this entire test is susceptible to being skewed by a downtrend, or any other random things..
while interesting, its' results will be statistically meaningless unless you do proper A/B testing
|
Very true.
If one month the traffic that comes stops signing up and the next month the traffic start signing up again. For no reason what so ever it will make the whole experiment rendered useless.
What's the odds of that then?