How do you deal with split testing ? (13)
I have a workflow problem that I cannot decide on.
When A/B testing landing pages. so I have lp1a lp1b for example. once you found a winner. do you create a new LP to test against the winner or 'overwrite' the looser with the winner and modify another element.
that sound minor, but in one case I end up with lots of LP... cpvlab get messy, but on the other case I loose the data for the previous test.
hope I make sense lol
another option would be something like 'Landing page genuis'...
i'd keep the data...the colors change in CPVLab when an LP is inactive...not that big of a deal to me.
thx, constantin, but still after 20-30 iterations, how people deal with all the different LP's ?
The only issue I have with having a lot of LPs is just finding the right one when I want to generate a new tracking link. Otherwise I don't see or deal with it, since I'm really only sending to a few at a time for a given offer and it doesn't show me stats for inactive LPs in Prosper202 at least.
I asked this question last week; there was some good discussion here.
thx guys, I'm not talking about any technical issue but more of a way of working and organizing things.
there's no reason to overwrite the loser with the winner which is maybe why you're not getting the answer you want. if you're doing something completely different like A vs. B you see which wins. lets say B does and you want a variation, you make a copy of B and change the element you need so you have B vs. B1 now and you switch traffic to B1 instead of A. No reason at all to touch A.
yeah, you see what I trying to understand, but I'm talking about small variations like headlines, cta etc..... I'm currently doing they way you are saying, but with lots of variations I feel its useless to have all those different copies when only a small element is tested... for example, with VWO or landing page genuis you only have one master document and its programaticly split tested.... again its not a that big deal, just trying to find a proper workflow. thx
Well at the point where you're testing many small things then I'd go with google website optimizer, which is apparently is now inside of google analytics.
http://support.google.com/analytics/...answer=2661700
Or there is http://landingpagegenius.com/ which is a prosper add-on for multivariate testing. If I had the spare cash I'd probably go with that to save time, but learning how to use analytics could be very handy in the future.
edit: well apparently the tester inside GA sucks balls now and doesn't do multivariate testing like GWO did. Damn.
it look like a lot of people are using LPG.
I'm going in....
LPG is pretty good -- we used it with Prosper for a while and it's actually very easy to do things.
I prefer using Optimizely these days as it's a lot, lot easier to set things up for more advanced landing pages than a one-page PPV pop, and it does all the more advanced statistics and changes for me (when I tested LPG there was no automatic pruning / optimization / confidence % / r2 value / etc.). In any case, LPG is definitely worth a try at the price since it's pretty cheap. 
I just started experimenting with LPG and it's made it a lot easier for me. I've been "in the game" for a long time, so I'm used to keeping it classic with the actual force split testing lol.
Home > General > Affiliate Marketing Forum