Where is the control for the experiment? What is the ROI on non-targeted ads?
For us we make a donation page. The only links that go there are from FB ads. We also use different refcodes for different ad groups AND the FB pixel passing in purchase value which is really good at optimizing FB's delivery to highest ROAS
FB is very very very good at this
I feel like people who question it or like don't understand basic attribution should do some more research before denigrating
There are other non-direct measurement, FB for instance allows you to upload offline purchases. Definitely not as clean especially if you're hitting across different channels. But a basic A/B holdout test is pretty good in that case too.
For us personally we cannot get anywhere close to similar ROI from other platforms like Snap that attempt to be FB let alone contextual targeting.
It probably works better for app instal ads but otherwise it's.
The cookie based ecosystem also doesn't add a ton of value, some studies say between 5-15%.
It's Facebook and Amazon and Google's 1P data that don't rely on cookies.