There's tons of content here and out there regarding how to create and maintain a profitable campaign, etc. That's all great, but we all know that we have to go through many failed campaigns to get there (right?).
I think its important to know when to cut and minimize your losses.
So the big question that leaves us is, when should you call a campaign a failure and start over with a new one?
I wrote an article about that topic here
http://www.charlesngo.com/losing-mon...g-or-to-pivot/
Usually you know, your gut will tell you. There's never an exact answer though. If you've tried a few banners/angles/landing pages and couldnt hit break even, then i'd say move on.
There has to be an answer to this based on statistical maths.
For each traffic source develop a strategy for culling ads and angles and stick with it. There are often campaigns that are unprofitable but that one 'spark' of brilliance or insight can turn the tides - and unfortunately this isn't something that statistics can account for.