which way is the better strategy in running the first round of tests:
Run a single test without much targeting to see the statistical data spread before targeting, or run a group of campaigns(30-35 US man, 36-40 US man, 30-35 US women, 36-40 US women, etc) to see which target work the best.
A group of campaigns targetting different group is always recommended, but I remember reading a post which speaks about the merits of the first method. Could not find it. Do you have any ideas about this?
Which traffic source?
Plenty of FIsh
I'd go for a spread of narrowly targeted campaigns every time. You can start optimising to the particular age group that way, and there's a lot less cross-referencing required to tell things like which creatives work for which agegroup.
^^ I would say narrowly targeted campaigns. It's way more cost-efficient since you're not burning money on testing irrelevant demo's