I'd love to see that data, because my current pet theory is that adtech applies the same morals internally that they do externally - i.e. everyone tries to scam each other, pretending their tools work.
The people buying Facebook ads know if they work. You can track customers through every click to the sale. You can experiment with $10 if you want. Try it.
I built a large print-on-demand apparel marketplace. Most of the sellers were driving traffic with facebook ads. It worked great. Unfortunately (for the sellers) so many people jumped into the game that they bid up the ad rates. Now Facebook makes most of the money in the tshirt business... but you can't say the system doesn't work.
> You can experiment with $10 if you want. Try it.
Guess what, I actually thought about doing that just "for science" - setting up some nonsense page, launching a $5-10 ad campaign, and seeing what happens. I.e. measuring the impulse response of the system, in a way.
To be clear, I source my distrust to adtech partly from personal experience as well. I worked at a company that also did social media marketing once, in particular it would run content marketing on Facebook for people. I had a first-hand look at "how the sausage is made", and my overall impression is: marketers having no clue about statistics dump some charts into a Word file and write a story about how that graph means things are great; customers having no more clue and no way to verify effectiveness believe that. Both sides are happy, and money changes hands.
Also funny you mention print-on-demand, for two reasons. One, I'm building a side project in this space right now. Two, print-on-demand apparel are the ads that are pissing me off the most right now on Facebook (in particular a certain company that's named after a game animal saying "hello" in Hawaiian).
