That's the key - you can't just "experiment" with paid acquisition. You have to spend a little to gather enough data on what actually works for your audience.

That's the key - you can't just "experiment" with paid acquisition. You have to spend a little to gather enough data on what actually works for your audience. It might take $1000 in failed ads before you "get" your audience, but you might quit after spending $100. But the $900 in supposedly "wasted" ad spend will help you understand enough that you can make 10x higher revenue. If you approach software development as an iterative activity, you should approach marketing the same way too. Expecting great results at first try is akin to expecting completely bug free software at first go without any tests.