I know a dozen companies who spend significant ($1m+) a year on Google or Facebook ads and performance is tracked obsessively.
I think it was the CEO at BMW who said that if he only showed you BMW advertising the day before you bought a car, they'd go out of business because no one would ever buy a BMW. The brand only works because they start showing you the ads when you're three years old, in hopes that you'll buy one when you're 35. And when they sponsor TED or whatever it's not because they think anyone there is going to go out and buy the car, it's to prove that they're a luxury brand by showing that they can afford to spend a lot of money on advertising that everyone knows isn't going to work.
If you look at brand vs direct response, you can see that it's roughly split among digital advertising.[1][2] But as Seth mentions a lot of people use direct marketing to try to do branding and so don't bother carefully measuring their funnel, and most non-digital advertising skews toward brand advertising because it's obviously harder (or impossible) to measure. And digital is less than a third of total advertising.
Also look at companies of the Uber/Handy sort, who spend huge on direct marketing purposely at a loss in hopes of later creating some sort of lock in through network effects. In that case it's part measuring and part gambling in a way that's not straightforward to classify.
[1] http://www.emarketer.com/Article/How-Much-Industries-Spendin...
[2] https://www.comscore.com/lat/Insights/Blog/On-Branding-Versu...
Most people who spend money on ads just want to be seen as the kind of people who spend money on ads, or else they just want to make their boss happy in the short term by telling them that they bought ads on Facebook or whatever. The chain of accountability is very diffuse, and most results don't really get tracked.