[ethanbond]

Except that's the fundamental issue of advertising? Attribution has always been wickedly difficult. It was supposed to be easy with Google/FB (click -> sale, duh), but now people are clicking on fewer and fewer ads, so Google/Facebook et al. are repositioning as "oh no, they saw our ads for x seconds, we definitely swayed them."

[disgruntledphd2]

The issue is that views definitely have an impact. However, especially on mobile, this impact is very difficult to measure. So most of the DR industry uses last click, even though they know its incorrect, because there is nothing better.

[ethanbond]

The impact is difficult to measure everywhere all the time. How do you think they measure impact of billboards? Radio ads? TV ads? All by the same hand-wavy "there is nothing better" types of metrics.

I didn't say they don't have an impact, I said it's obscenely difficult to gauge attribution and online ads weren't the panacea they were supposed to be because it turns out people don't click ads. Despite this, advertisers are still advertising. An increase in ad spending is not any indicator that ads are working well because no one has ever known how well they worked.

[disgruntledphd2]

Definitely. This is a part of my professional life, so I get exposed to these issues all the time. TV, in particular, tends to be measured in very strange ways by DR advertisers because they appear to believe that advertising was invented along with the click, by Google (really Overture, but hey).

To be fair to offline, you can get a pretty good read on store sales as a function of advertising if you do geographical splits. Its modelled, and statistical, but there's definitely nothing wrong with that.

I think the online problem is harder because its so easy to measure clicks that people focus entirely on them without considering other approaches, whereas with offline its hard to get any measure, so people are willing to try different things.