these are the stats for an adult campaign im running… you may have read my follow along in the "follow along" section
banner1 has about 27k impressions with a .18 CTR… probably going to delete this one right now
banner5 I already deleted, I felt like that was pretty obvious
banner6 has 47k impressions and .27 CTR
banner2 has about 46k impressions with a .24 CTR
banner4 has 32k impressions and a .3 CTR
banner3 has 22k impressions and a .3 CTR
The reason why i get so confused about went to cut ads is that when i went to the a4d meetup there was a whole presentation about statistical significance, so I get really paranoid about trimming fat
should I keep banner 6 and 4? they have small roi’s and banner6 is barely even profitable…
How much $ have you spent on all these banners? Around $200+ each?
200$? on each banner? Nah just like 10-12$ haha, that is why im asking if it would be smart to start trimming the fat yet, since the conversions are still low
I tend to do what you do.. i rarley send a lot of traffic to each banner.. usually take out the bad ones from the beginning heheh
I think a post on significance across different traffic sources would be very useful for people around here (including me of course!) Like on sources such as POF I’ve heard you can make a decision off a few thousand impressions because it stays relatively constant.
What about PPV?FB? To get true statistical significance we’d have to spend A LOT on non performers.
You could also pause them, if you have a smaller budget. Then pick the two you think are winners and run them.
Pick the winner of those two and run against a paused ad, repeat until you get statistical sig.
I think i read something wrong then lol, eitherway with that said, i would only scrap banner5, and keep running the others, data changes, so don’t trim too early.
To give you a quick answer to what your asking: one thing to keep in mind like james was saying is never to trim too early on ANY traffic source.
"Like on sources such as POF I’ve heard you can make a decision off a few thousand impressions because it stays relatively constant. "
This may be true, but it’ll never be true all the time, how could it? There’s way too many variables. So to get true stats on your campaigns you never want to cut anything off too early. You don’t have to spend $200 on poor performers, but you don’t want to cut something just because its $2.00 in the red.
If your budget it small, just start small and move from there and keep the winners off the bat to build up the budget.
The Article Published IN 06-20-2011 12:45 AM