News for hackers. Stuff that matters.
Re:Apple’s ad business set to boom on the back of its own an... by whimsicalism on Monday October 03, 2022 @02:44PM (#33071894)
> doesn't necessarily require pervasive tracking

I disagree. Incrementality studies (which measure ROI) as advertisers want them are basically impossible with ATT. You need to be able to pass an ID between apps.

Re:Apple’s ad business set to boom on the back of its own an... by clairity on Monday October 03, 2022 @03:01PM (#33072111)
yes, but you haven't shown that that translates into more precise and accurate ROI. marketers and advertisers believe it should, but there's no solid proof. that's because markets (and any human endeavor) is complicated beyond our ability to model (and solve) it deterministically. attribution models (such as incrementality studies) can sometimes give you clues, but can't really tell you why any given person bought something with any certainty. it's the old adage of half of advertising dollars are wasted, but you don't know which half.
Re:Apple’s ad business set to boom on the back of its own an... by whimsicalism on Monday October 03, 2022 @05:38PM (#33074053)
> there's no solid proof

I think modern incrementality studies give solid proof of the value of an ad on the basis of a good model of how the world works. Of course, models can be wrong - it could turn out that solipsism is true, physics is false, and the world outside of your own mind is a figment of your imagination!

That the world is complex and models are inherently wrong does not mean that nothing of value to businesses was lost with ATT.

Re:Apple’s ad business set to boom on the back of its own an... by clairity on Tuesday October 04, 2022 @02:27PM (#33084619)
no, there's a belief of value (and fomo), but not concrete proof. there's little correlation between the price of an ad and the value of attribution beyond the value of the ad itself.
Re:Apple’s ad business set to boom on the back of its own an... by whimsicalism on Tuesday October 04, 2022 @11:02PM (#33090598)
Incrementality studies provide concrete proof of how your ad impacts behavior. How you value that behavior change is up to the business and reflected in the price they are willing to pay.
Re:Apple’s ad business set to boom on the back of its own an... by disgruntledphd2 on Monday October 03, 2022 @03:45PM (#33072676)
Before ATT happened, FB has a tool that would run experiments to assess incrementality of your advertising. It's possible, but there are a bunch of privacy trade-offs..
Re:Apple’s ad business set to boom on the back of its own an... by clairity on Monday October 03, 2022 @04:54PM (#33073563)
right, but again, those are very likely probabalistic, population-level models that make assumptions about how to attribute credit--does it all go to the first view/click? how likely is the first view/click really the first view/click? do you instead apportion credit across clicks/views? how? it's somewhat useful at a population level, but not at all at an individual level, especially not for the tradeoff in privacy, anonymity, and autonomy.

but the kicker is, is it better than just doing studies without the more invasive attribution data, especially in relation to the higher price and market consolidation? very unlikely. ad monopolization means more of the value in the value chain goes to the monopolist regardless of the proportion of value they provide in the chain.

Re:Apple’s ad business set to boom on the back of its own an... by whimsicalism on Monday October 03, 2022 @05:39PM (#33074068)
> is it better than just doing studies without the more invasive attribution data,

Absolutely, as the controlled incrementality study is impossible without either attribution or some group-based approximation of attribution (ie. federated cohorts, etc.)

Re:Apple’s ad business set to boom on the back of its own an... by disgruntledphd2 on Tuesday October 04, 2022 @04:02AM (#33077642)
> right, but again, those are very likely probabalistic, population-level models that make assumptions about how to attribute credit--does it all go to the first view/click? how likely is the first view/click really the first view/click? do you instead apportion credit across clicks/views? how? it's somewhat useful at a population level, but not at all at an individual level, especially not for the tradeoff in privacy, anonymity, and autonomy.

So the idea is that you run the experiment, and this can then help you understand where you should attribute value, as you know that the only difference between the two groups was the exposure to FB ads.

Now, to be fair, unless FB is most of your spend, you still have a bunch of problems, but they'll wash out equally across conditions (theoretically, at least) so the estimate should be good.