The only reason (yes I asked) Facebook continued to allow these studies was because Google did even worse.
I wonder if all this stuff ever got published.
Never heard anything in the press, so I am guessing Facebook eventually pulled the plug. I wonder if I saved the drafts tho :>
https://freakonomics.com/podcast/advertising-part-2/
One thing I was hoping they'd explore was whether any of the recent companies that participated in the fb ad boycott realized afterward that the advertising dollars were not really working for them.
Apparently - I did a quick search - conversion rate at facebook is said to be on average up to 9% (in marketing materials) or around 4-5%.
In this study, the authors show conversion rate consistently sub 1%, which is however not the point of the article. Rather, they also show that the lift in conversion (so the actual impact of the ad) as calculated by observational measures usually over-estimates the actual lift (sometimes significantly). Observational measures were those used by advertisers on facebook - maybe even now?
The more snarky comments from the presentation are missing from the draft, of course.
I don't understand - a "slight" error and 60-80% is "trivial" for a "short" period of two years? Is it just me or is this language excessively apologetic? Or is it meant to be ironic?
How long a company can get away with that depends on their moats, competitive landscape, and what alternatives exist. In the case of FB, obviously there's a huge moat and a semi monopoly.
There is some competition from Twitter and Snapchat, but really FB acquiring Instagram doubled their market cap, because that is the only other social platform that has truly reached their scale, and also has an effective, if not more effective, ad platform.
This is also pretty typical even at smaller companies. Different people have different "masters". The product manager is rightfully trying to build the best product for their customers - the advertisers. But the upper management has bonuses tied to revenue usually through stock compensation, so their interest is in their own pockets. The product manager who correctly points out a real issue, is then seen as a negative, rather than someone doing a fantastic job.
In my opinion, anyone involved in such large-scale data collection and analysis should acknowledge the inevitability of error and provide disclaimers about potential sources of error.
The advertisement companies know this very well, but they also know that the specific, well-aimed ads make a small fraction of their revenue. So they have a very direct interest in using the success metrics from the specific ads to sell their platform to a much wider range of customers, for whom it will never pay off.
There's also considerable FUD[0] involved in the decision process. Sure, you don't see your ads paying off directly, but there must be that legendary brand awareness building up in the background. What if you decide to pull the plug on ads, and in a year your company's sales drop by 20%? Do you want to risk your entire career over a choice of saving your employer's money that isn't your own money anyway?
Also a huge chunk of advertisers are clueless people that decided to try ads for the first time. They don't know what to expect in terms of payoff, they don't know to make specific problem-oriented ads (and whether their case is even a problem-oriented case). So they spend some ad dollars and move on. But due to the scale of the market, it still adds up to a healthy profit for the platform.
[0] https://en.wikipedia.org/wiki/Fear,_uncertainty,_and_doubt
Once enough money has entered the picture, people become very flexible.
https://which-50.com/how-uber-learned-it-had-a-huge-ad-fraud...
I don't think they turned a page yet.
Rince and repeat; real definition of 'serial entrepreneur'.
Google, as it exists now, is the natural outcome of this process.
