Google uses their panopticon to show your ad to a user who is just about to buy your product and then claim a conversion. So the stats look like Google is getting you thousands of conversions, when they only actually got a hundred people to look at your product who weren't already interested.
This kind of bullshit was not possible in legacy advertisement. A billboard cannot change itself to always be showing an ad that can be claimed as a conversion to every single user.
The newspaper ads could not change to ensure that you saw an ad that matched what you were about to purchase.
>There are a lot of statisticians working in this area - or there were 20 years ago
Weren't those people the exact ones who came up with "Half of ad spend is wasted but we don't know which half"?
Targeted and online tracking based advertising has fundamental information asymmetry problems that fuck over everyone but Google and Meta.
yes a lot of ad spend is wasted but they can prove what was still useful enough to be worth spending despite the waste and which was not.