Re:The $300B Google-Meta advertising duopoly is under attack
by CryptoBanker on Sunday September 18, 2022 @06:48PM
(#32892222)
That experiment simply means their ads don’t work
Re:The $300B Google-Meta advertising duopoly is under attack
by asdff on Monday September 19, 2022 @03:22AM
(#32895319)
How do you know your ad even worked and you didn't just correlate performance of your ad campaign with some confounding factor(s) you will never be able to account for? you aren't testing for a mechanism here, just whether certain other metrics changed favorably over the campaign period and you make assumptions about the cause. you have no idea if the ad converted or not. how may times are you repeating your ab testing and with what sample size? are you running simulations too so you understand what a successful campaign would even look like based on your sample size? are you underpowered or underdetermined? probably only the most careful advertisers are thinking about these things, and I really wonder how significant the results look after doing these sorts of things.
Re:The $300B Google-Meta advertising duopoly is under attack
by huitzitziltzin on Sunday September 18, 2022 @08:26PM
(#32892849)
Very true. That’s part of my point. It’s a hard question! It is not correct to assume that advertising works. We don’t know.
Re:The $300B Google-Meta advertising duopoly is under attack
by najsudhsis on Sunday September 18, 2022 @08:08PM
(#32892730)
its a simple question answered by sociologists not economists.
advertising creates a barrier to enter a market, not the market.
when you spend on ads, you are not creating new market for you, you are just unlocking some of the reserve market the established leaders secure via misinformation and brand brainwashing.
Re:The $300B Google-Meta advertising duopoly is under attack
by huitzitziltzin on Sunday September 18, 2022 @08:25PM
(#32892846)
We have been considering that possibility in the literature on the economics of advertising since the 1950’s or 1960’s, if not earlier. If you want a survey of this and other early ideas in the literature, see Kyle Bagwell’s contribution to the handbook of industrial organization, vol 3 (2007 or so).
Re:The $300B Google-Meta advertising duopoly is under attack
by echelon on Sunday September 18, 2022 @11:13PM
(#32893879)
I know this is a terrible imposition, but would you mind summarizing the conclusions of this comment and your previous comment with regard to the measurements done in literature?
This is hard to follow for those of us not in academia with access to journals. Or time to figure out how to hunt these sources down.
I feel like your comments contain pointers to deep quantitative insights.
One place to start on the difficulties is an excellent paper by Lewis and Rao from the Quarterly Journal of Economics in 2015. The original title was something like: “on the impossibility of measuring the returns to advertising.” Just getting enough data to reject the hypothesis “this ad campaign did nothing” is extremely challenging. Lewis is a great producer of research on this topic and has written papers based on his experience at Yahoo.
Another great paper is by Blake, Nosko and Tadelis from Econométrica in… 2014? They turned off all of eBay’s keyword search ads in some markets as an experiment. They found that they maintained about 95%+ of their business while saving $50 eBay million dollars or so.
Beware confident assertions from people in the ad industry that ads clearly work. It is not so obvious that they do. This isn’t to say they don’t work! But it’s a challenging scientific question.