Removing creatives based on confidence levels (3)
I have been reading older threads and it always seems that people advocate testing with a lot of impressions so you don't make decisions with insignificant data.
I've done PPC for the most of my affiliate career so I was wondering if somebody can tell me if, for example http://splittester.com/, works for deciding whether there's enough data to pick a winner from +2 creatives? I mean all that calculator does is calculate statistical significance so at least I can't think of a reason why this same model wouldn't apply to banners. But please do correct me if I'm wrong 
yes that tool works fine.
but in my experience you often dont need to wait till the 95% confidence level! Settling on 85% requires a lot less impressions.
And make sure you factor in the number of variables your testing, ie: 3 creatives, 2 landers, 2 offers = 12 variables
Yeah, I should use ad tracking in 202 to get some numbers on how ads convert. At least in SiteScout you can't define link per ad so I would have to do yet another processing layer at my rotator that would match up SiteScout ad id to 202 ad id. Another way is with custom variables but we'll see which one fits in the overall system better.
Home > Paid Traffic Sources > Display / Media Buying