The two things do not seem to me as contrasting, there is an industry gathering the data and selling them (with promises and snake-oil) and other people buying those data.
Whether what is sold is accurate or not is another thing, as long as it is perceived as "good enough" by the buyers it will be bought anyway.
I don't think that many companies (buyers) have the guts to stop for - say - 6 months all advertising on a given channel and see what happens.
Again, take the data buying and selling and set it aside. That's kind of like a whole separate thing.
Even without data brokers, there is tons of fraud and snake-oil. When you are comparing advertising success, website traffic, email delivery, product usage, etc - actually getting consistent user data between them is a nightmare. Everyone has incentive to pump up their own numbers. And oftentimes the reporting you are engaging in is so complicated that a lot of companies rely on consultants to do all of the work.
> I don't think that many companies (buyers) have the guts to stop for - say - 6 months all advertising on a given channel and see what happens.
Our company had a broken ad run on LinkedIn for 6 months. It cost us over $100k. The manager's response was "Can we turn them on again until we fix them? I need to spend the budget or I will lose it."
Yes, we are saying exactly the same things, though in a slightly different way.
The results of this (or that) advertising are either non-measurable or poorly measured, and the way advertising budgets are calculated (and spent) is - to say the least - perverse.
Whether what is sold is accurate or not is another thing, as long as it is perceived as "good enough" by the buyers it will be bought anyway.
I don't think that many companies (buyers) have the guts to stop for - say - 6 months all advertising on a given channel and see what happens.