The first thing one has to face when considering such experiments is that they fundamentally have an inherent cost. Because in order to get that measurement accurate, quite a lot of your ad impressions have to be withheld. This is a cost, a cost of the lost opportunity. (provided that the actual impact of the ad campaign is positive, that is).
So, as an advertiser, you end up coming up with some meta-strategy: when to perform those experiments and when to forgo them and rely on your educated guess.
Google, Facebook and alike developed support for performing such experiments throughout the last decade. But this meta-strategy kind of judgment is firmly in the department of advertisers.
Certainly not a new dot com bubble as the title suggests, and the article even mentions Randall Lewis, so I guess it's just the style of journalism these days...
So, as an advertiser, you end up coming up with some meta-strategy: when to perform those experiments and when to forgo them and rely on your educated guess.
Google, Facebook and alike developed support for performing such experiments throughout the last decade. But this meta-strategy kind of judgment is firmly in the department of advertisers.
Certainly not a new dot com bubble as the title suggests, and the article even mentions Randall Lewis, so I guess it's just the style of journalism these days...