Sometimes it's really hard to measure the effects of online ads

Or as Jesse Frederik and Maurits Martijn named it The new dot com bubble is here: it’s called online advertising .

It goes in length describing the case where eBay stopped paying $20mln annually for buying brand-name search ads from Google and did not see any visitors decrease. Then finishes with 13 studies of Facebook ad campaigns, in 10 of which the vast majority of sales increases were attributed to selection effect (e.g. customers were already looking for Nike shoes but went to through a Google ad).

Advertising rationally, the way it’s described in economic textbooks, is unattainable. Then how do advertisers know what they ought to pay for ads?
“Yeah, basically they don’t know,” Lewis said in one of those throw-away clauses that kept running through my head for days after.

The article stops there, but I’ll have to argue that most of the examples apply to big brand companies. While a startup with a proven niche and known unit economics can and should still benefit from ads. Just not the brand-name ones, but the ones that help a user to discover their service (e.g. ‘cooked meals delivery’).