I finally learned a good things last week reading an article on a marketing blog saying that no matter how much money you lose on an initial campaign, you shouldn't think about this as "losing money", you should rather think about it as "spending money to get some data about that specific campaign" and the best thing is that the first campaign you create is usually the worst of all and from there on, you'll only get it better by optimizing it. That's why I decided to spend some more money on PPC campaigns even though I "lose money" initially, but in the long term, by optimizing them, I can actually make them profitable.
The same principle applies to split testing and other tests for improving conversions.
Of course, this is just theory for me, because I'm still in the phase of losing money on the initial campaign, but hopefully very soon I'll get enough data to allow me to optimize at least one campaign to make it profitable