Quote:
Originally Posted by MisterPeabody
I've actually begun "split testing" or "A/B testing" and it's a simple script Woj wrote up for me; it diverts every other hit to "B". So say i get 1000 hits from YouPorn (where Mia is on right now). 500 go to A, 500 go to B. Every other hit.
|
Problem I would see would be this: How do I know if "A" or "B" may or may not have gotten more "good" traffic and/or surfers more likely to join?
I wouldn't really. It's a 50/50 chance no matter what in that scenario. Maybe today 75% of the 50% that went to "A" were 16 year olds surfing porn. Maybe tomorrow that percentage just happens to be middle aged guys who love big tits and have credit cards.
I wouldn't really have any way to know that both "A" and "B" are getting exactly the same quality of traffic. Even if they came from only one source.
So I'm thinking the only real test is to just change it. Give it enough days to make sure it has a shot at real convert-able traffic and see what happens.
Any other way of "testing" is still susceptible to chance.