Jan 26, 2023·edited Jan 26, 2023Liked by Noah Smith
I just don't see how one could have a good reason to pursue this lawsuit against google without pursuing apple. Apple is openly and blatantly using it's market power in hardware to extract software rents. There is no question there are network effects and it would probably be a national security plus not minus.
It all makes me wonder if this isn't, in part, driven by a certain resentment in the media sphere that google has eaten into traditional newspaper profits. That's probably just paranoid thinking but it just feels nutty.
I can't compare the overall consumer impact but both the legal case and solution seem much clearer for apple. Bank a win by forcing apple to allow alternative app stores and then think about the less clear cut cases.
As someone who works in the digital marketing world I can tell you that Google has bigger issues, it used to be that we’d get about 1,000 bidders a week from paid search and now even with more spend (their prices have been jacked up) we’re down to about 300 bidders a week even with the same number of UVs, which suggests that their search terms and keywords are not as effective at attracting actual customers as it used to be, plus I’m pretty sure that anyone using Google lately has noticed how much poorer the search function is over the last couple of years
On top of that ChatGPT is gonna eat into search, just this week I was googling some answers about some changes I wanted to make to SQL code and wasn’t finding the answers I needed, back to that search issue it’s not as good as interpreting your questions to deliver the results you want as it used to be
So I decided to try ChatGPT and started with a Google search type thing and got a Google search type answer (but straight away got the closest I got from Google after about 20 minutes of attempts) but then I just dropped on the code with a question for the change I wanted and BOOM it just updated the code for me, there was one issue with the code (it gave me what I wanted but made me lose one thing I already had and needed) so I just typed in ‘but I still need X’ and BOOM it just delivered the updated code
Ppl are looking at ChatGPT all wrong, it’s not about AI conversations or whatever, it’s optimised search and will take a huge chunk of market share off Google, I know I’ll go to ChatGPT first now when looking for any technical help or information and I’m sure they’ll find a way to monetise it soon
I know it’s more outside the economist’s purview, but I would like to have seen more consideration given to whether and when the power struggle aspect of anti-trust is valid. Google being able to control so much of the pathways people use to access information on the internet between search and ad dominance, along with the prominence of YouTube, seems prima facie to be a valid, serious concern. Maybe breaking them up still is not the answer in this case, but it seems perfectly reasonable for the populace to be worried about a large corporation having outsized, unbalance power like that, completely outside of any economic implications.
‘Why would advertisers and websites together pay 30% of ad dollars to Google? There are two possible reasons. Possibility 1 is that Google’s tool’
That used to be the case but not anymore, see my previous reply, the quality of Google search results both SEO & SEM have dropped off considerably, they still deliver the UVs but they seem more effective at attracting bots than actual ppl when u look at conversion
One other thing I’d like regulator’s to look at is the reporting tools like Google Analytics and Adobe Analytics
Just one thing you notice is that when traffic drops you also see that average time per visit drops (and same with increases) now there is no reason that the average visit time should go up when traffic goes up or down when traffic goes down, think about it, if 2 ppl visit or 20 ppl visit it shouldn’t make any difference to the average length of those visits (and certainly not when it’s 1.5 million or a million) but as digital marketing analysts we really only have these 2 tools to go on and they do usually pretty much agree, but this issue makes me question many other of their numbers (things like getting 1.5 million UVs per month but 14 million UVs per year, I just don’t buy that we’re getting that few return UVs month to month, or that 2 out of every 3 Australians visited our eCommerce site last year) but because Adobe and GA generally agree and the numbers are high nobody has any motivation to question the numbers being delivered even though any thinking person has to question how real they really are
The iPod being a more lucrative (and superior product) to the MP3 players that came before them is a nice historical irony with windows being a more lucrative take on the Mac operating system
1) AI is overrated. Name a profitable AI application aside from my Zojirushi rice cooker. What percentage of Google's revenues does it represent? AI is just window dressing. It's self driving cars. When Google released an actually useful AI system, Alphafold, it took researchers a few months to release an open source version. When AI actually starts to make a difference, we'll see that Google was nearly irrelevant.
2) If you are looking a monopoly power, one important issue is whether Google demands exclusive rights to place ads on a site. I'm guessing they do. That precludes the development of more focused advertising markets. A lot of them were developing 20 years ago but Google ate them all.
3) If you break up Google, follow the AT&T model. Break out the indexing and ranking component, the web crawler, from the search engines that provide the actual search results to user queries. This would open a market for more focused search markets. Google search is garbage these days, but it's the only game in town. A lot of people will go for natural language query and response while others will go for more sophisticated search. It would be nice to have some options, and the advertisers would like having a segmented market.
"So it probably was with Microsoft in the 90s and early 00s, when people thought — wrongly, it turned out — that control of desktop operating systems made that company an all-conquering empire."
Not sure of this statement as Apple, current market cap $2.25T and Microsoft, $1.8T, dominate mobile and desktop operating systems today.
The best historical example that is analogous to Google is that of AT&T. The Judge Greene decision in 1983 that broke up AT&T's monopoly on voice and data, absent MCI's competition for long-distance that began in 1973, ultimately led the way to the wide-spread adoption of the internet in the early 1990s, as telecommunications equipment and services became subject to market competition which led to lower prices and better services. You did not have to employ an AT&T branded or "approved" devices to connect to the local phone network anymore, so independent manufacturers of modems, as an example, such as Hayes and Motorola, could innovate and compete with the lagging technology of Ma Bell. AT&T was relentlessly focused on their profit center of long distance calling and sales of automated switches to corporations to reduce labor costs, as well as in their own network. In the 1940s telephone switchboard operators numbered as high as 340,000 persons in the U.S. alone.
Google is similarly situated today to AT&T in the early '80s. A market cap of $1.24T and a stranglehold on search and ads. Apart from Ticketmaster, they are a reasonable target of anti-monopoly legislation or executive action.
Google will never be broken up because it is too useful to the national security state. Google collects all the information that governments can't, because of constitutional protections, and then turns it over when asked to do so. And this doesn't even consider their role in shaping what we are allowed to see, in the guise of protecting the "cognitive infrastructure." It's the same reason the Chinese government prefers monopolies.
I think the issue is a bit simpler, especially once you add in allegations of Bell/Bernanke fraud (collecting $$$ selling a display ad, telling publishers the auction closed at $$ and underpaying them by $, using the $ withheld for seemingly anticompetitive actions in another layer of the display market).
Also, executives involved should maybe search for criminal defense lawyers, just in case, as it appears the allegations point out an intentional criminal fraud.
Right now Google is the auction house and they also want you to let them bid for you. So they are the auction house, and they are your personal bidder. But they are also your competitors personal bidder. And all your other competitors personal bidders.
This actually makes sense despite seeming absurd, they have the data, and you almost always WILL get better results if you turn over your bidding to them to do. Or at least you will now vs manual bidding now, when some of your competition are letting Google run their account.
But if Google runs the auction, and runs all the bids for all auctions (after your competitors also convert to letting Google do it all), what will they optimize the total end result for?
Would it ever be lower prices? They can't save everyone money without lowering revenue, and "monopoly pricing" along a demand curve seems to be exactly what this setup could allow if Google gets to control all pricing and data.
But I agree we do have to worry about what it means to have Google releasing that much data to other vendors (it is not private anymore, and can be assumed to just be out there once duplicated that many times). A semi-strong monopoly power being used to collect extra profit might not be worse than the alternatives, if the alternatives require all auction data and other data being released to a whole market of vendors replicating their bidding automation.
My understanding is that in antitrust at present there are only certain types of reasons that are recognized as legitimate reasons to take action. Now, again as a layperson, it seems what counts as a legitimate reason shifts around and isn't totally set in stone (the consumer welfare paradigm started by Bork is popular among free market types, Tim Wu and Lina Khan want to shift the Overton window here)
But it strikes me as not on the table presently to use anti trust to do industrial policy by the courts. Like it may be true that if Google invested in this stream of innovation rather than just ads it'd lead to cooler AI which would be great in a contest in China - but is that *reason* that could actually be *given* in the anti trust proceedings? I think that would be a pretty bad way to proceed with antitrust because while being competitive against China is good, it doesn't provide anything close to an objective standard to assess anti trust and just seems very far away from any rationale of anti trust.
Don't we have bigger fish to fry? I don't feel that I'm being screwed over by Google and Apple. I think they have improved our lives. We have plenty of competition at the platform level (Google vs. Apple vs. Microsoft vs. Amazon). Security is better with large ecosystems. People trust Apple and Google, so they are incented to regulate themselves in order to keep that trust. Meta, not so much. The technology itself will force it to evolve to stay competitive (see the good comment on ChatGPT here and think about how ad blockers will force Google to innovate to find new revenue streams).
I just don't see how one could have a good reason to pursue this lawsuit against google without pursuing apple. Apple is openly and blatantly using it's market power in hardware to extract software rents. There is no question there are network effects and it would probably be a national security plus not minus.
It all makes me wonder if this isn't, in part, driven by a certain resentment in the media sphere that google has eaten into traditional newspaper profits. That's probably just paranoid thinking but it just feels nutty.
I can't compare the overall consumer impact but both the legal case and solution seem much clearer for apple. Bank a win by forcing apple to allow alternative app stores and then think about the less clear cut cases.
As someone who works in the digital marketing world I can tell you that Google has bigger issues, it used to be that we’d get about 1,000 bidders a week from paid search and now even with more spend (their prices have been jacked up) we’re down to about 300 bidders a week even with the same number of UVs, which suggests that their search terms and keywords are not as effective at attracting actual customers as it used to be, plus I’m pretty sure that anyone using Google lately has noticed how much poorer the search function is over the last couple of years
On top of that ChatGPT is gonna eat into search, just this week I was googling some answers about some changes I wanted to make to SQL code and wasn’t finding the answers I needed, back to that search issue it’s not as good as interpreting your questions to deliver the results you want as it used to be
So I decided to try ChatGPT and started with a Google search type thing and got a Google search type answer (but straight away got the closest I got from Google after about 20 minutes of attempts) but then I just dropped on the code with a question for the change I wanted and BOOM it just updated the code for me, there was one issue with the code (it gave me what I wanted but made me lose one thing I already had and needed) so I just typed in ‘but I still need X’ and BOOM it just delivered the updated code
Ppl are looking at ChatGPT all wrong, it’s not about AI conversations or whatever, it’s optimised search and will take a huge chunk of market share off Google, I know I’ll go to ChatGPT first now when looking for any technical help or information and I’m sure they’ll find a way to monetise it soon
Display Ads are a foster child in Google:
For Q3’22 the revenue was:
Search: $39.5B
YouTube: $7B
Display Network: $7.9B
Total Ads: $54.5B
Total Alphabet revenue: $69B
So Display Ads revenue is 14.5% of Google Ads revenue or 11.5% of total Alphabet revenue.
So “smaller Google” here would mean like 5% smaller.
I know it’s more outside the economist’s purview, but I would like to have seen more consideration given to whether and when the power struggle aspect of anti-trust is valid. Google being able to control so much of the pathways people use to access information on the internet between search and ad dominance, along with the prominence of YouTube, seems prima facie to be a valid, serious concern. Maybe breaking them up still is not the answer in this case, but it seems perfectly reasonable for the populace to be worried about a large corporation having outsized, unbalance power like that, completely outside of any economic implications.
‘Why would advertisers and websites together pay 30% of ad dollars to Google? There are two possible reasons. Possibility 1 is that Google’s tool’
That used to be the case but not anymore, see my previous reply, the quality of Google search results both SEO & SEM have dropped off considerably, they still deliver the UVs but they seem more effective at attracting bots than actual ppl when u look at conversion
One other thing I’d like regulator’s to look at is the reporting tools like Google Analytics and Adobe Analytics
Just one thing you notice is that when traffic drops you also see that average time per visit drops (and same with increases) now there is no reason that the average visit time should go up when traffic goes up or down when traffic goes down, think about it, if 2 ppl visit or 20 ppl visit it shouldn’t make any difference to the average length of those visits (and certainly not when it’s 1.5 million or a million) but as digital marketing analysts we really only have these 2 tools to go on and they do usually pretty much agree, but this issue makes me question many other of their numbers (things like getting 1.5 million UVs per month but 14 million UVs per year, I just don’t buy that we’re getting that few return UVs month to month, or that 2 out of every 3 Australians visited our eCommerce site last year) but because Adobe and GA generally agree and the numbers are high nobody has any motivation to question the numbers being delivered even though any thinking person has to question how real they really are
The iPod being a more lucrative (and superior product) to the MP3 players that came before them is a nice historical irony with windows being a more lucrative take on the Mac operating system
1) AI is overrated. Name a profitable AI application aside from my Zojirushi rice cooker. What percentage of Google's revenues does it represent? AI is just window dressing. It's self driving cars. When Google released an actually useful AI system, Alphafold, it took researchers a few months to release an open source version. When AI actually starts to make a difference, we'll see that Google was nearly irrelevant.
2) If you are looking a monopoly power, one important issue is whether Google demands exclusive rights to place ads on a site. I'm guessing they do. That precludes the development of more focused advertising markets. A lot of them were developing 20 years ago but Google ate them all.
3) If you break up Google, follow the AT&T model. Break out the indexing and ranking component, the web crawler, from the search engines that provide the actual search results to user queries. This would open a market for more focused search markets. Google search is garbage these days, but it's the only game in town. A lot of people will go for natural language query and response while others will go for more sophisticated search. It would be nice to have some options, and the advertisers would like having a segmented market.
Google also prevents developers from mixing their services with a competitor's service. For example, the terms of service prevent using Google Search in conjunction with a Mapbox map. https://www.congress.gov/117/meeting/house/111247/witnesses/HHRG-117-JU05-Wstate-GundersenE-20210225.pdf
There are similar rules in place when using Google Cloud Services.
A Microsoft-like smack down, like Noah mentioned, in his post is probably the answer here.
Let me just be the first to performatively accuse you of being a neoliberal shill without actually reading the article.
Now that we've got THAT out of the way...
Bell Labs a stream of important innovations.
Google, what are their fundamental innovations?
Bell Labs innovations freely licensed at low prices.
Google pricing structure for their innovations?
Break them up.
"So it probably was with Microsoft in the 90s and early 00s, when people thought — wrongly, it turned out — that control of desktop operating systems made that company an all-conquering empire."
Not sure of this statement as Apple, current market cap $2.25T and Microsoft, $1.8T, dominate mobile and desktop operating systems today.
The best historical example that is analogous to Google is that of AT&T. The Judge Greene decision in 1983 that broke up AT&T's monopoly on voice and data, absent MCI's competition for long-distance that began in 1973, ultimately led the way to the wide-spread adoption of the internet in the early 1990s, as telecommunications equipment and services became subject to market competition which led to lower prices and better services. You did not have to employ an AT&T branded or "approved" devices to connect to the local phone network anymore, so independent manufacturers of modems, as an example, such as Hayes and Motorola, could innovate and compete with the lagging technology of Ma Bell. AT&T was relentlessly focused on their profit center of long distance calling and sales of automated switches to corporations to reduce labor costs, as well as in their own network. In the 1940s telephone switchboard operators numbered as high as 340,000 persons in the U.S. alone.
Google is similarly situated today to AT&T in the early '80s. A market cap of $1.24T and a stranglehold on search and ads. Apart from Ticketmaster, they are a reasonable target of anti-monopoly legislation or executive action.
EDIT to add:
See https://www.washingtonpost.com/archive/politics/1983/07/09/judge-orders-changes-in-bell-divestiture-plan/ca16d897-5414-425c-8d55-12b5b9eafb15/
See also https://ethw.org/Telephone_Operators
See also https://money.cnn.com/magazines/fortune/fortune500_archive/full/1980/
AT&T was only in 19th place in revenue in 1980.
See also https://money.cnn.com/magazines/fortune/fortune500_archive/full/1980/
Google will never be broken up because it is too useful to the national security state. Google collects all the information that governments can't, because of constitutional protections, and then turns it over when asked to do so. And this doesn't even consider their role in shaping what we are allowed to see, in the guise of protecting the "cognitive infrastructure." It's the same reason the Chinese government prefers monopolies.
I think the issue is a bit simpler, especially once you add in allegations of Bell/Bernanke fraud (collecting $$$ selling a display ad, telling publishers the auction closed at $$ and underpaying them by $, using the $ withheld for seemingly anticompetitive actions in another layer of the display market).
Also, executives involved should maybe search for criminal defense lawyers, just in case, as it appears the allegations point out an intentional criminal fraud.
Right now Google is the auction house and they also want you to let them bid for you. So they are the auction house, and they are your personal bidder. But they are also your competitors personal bidder. And all your other competitors personal bidders.
This actually makes sense despite seeming absurd, they have the data, and you almost always WILL get better results if you turn over your bidding to them to do. Or at least you will now vs manual bidding now, when some of your competition are letting Google run their account.
But if Google runs the auction, and runs all the bids for all auctions (after your competitors also convert to letting Google do it all), what will they optimize the total end result for?
Would it ever be lower prices? They can't save everyone money without lowering revenue, and "monopoly pricing" along a demand curve seems to be exactly what this setup could allow if Google gets to control all pricing and data.
But I agree we do have to worry about what it means to have Google releasing that much data to other vendors (it is not private anymore, and can be assumed to just be out there once duplicated that many times). A semi-strong monopoly power being used to collect extra profit might not be worse than the alternatives, if the alternatives require all auction data and other data being released to a whole market of vendors replicating their bidding automation.
My understanding is that in antitrust at present there are only certain types of reasons that are recognized as legitimate reasons to take action. Now, again as a layperson, it seems what counts as a legitimate reason shifts around and isn't totally set in stone (the consumer welfare paradigm started by Bork is popular among free market types, Tim Wu and Lina Khan want to shift the Overton window here)
But it strikes me as not on the table presently to use anti trust to do industrial policy by the courts. Like it may be true that if Google invested in this stream of innovation rather than just ads it'd lead to cooler AI which would be great in a contest in China - but is that *reason* that could actually be *given* in the anti trust proceedings? I think that would be a pretty bad way to proceed with antitrust because while being competitive against China is good, it doesn't provide anything close to an objective standard to assess anti trust and just seems very far away from any rationale of anti trust.
Don't we have bigger fish to fry? I don't feel that I'm being screwed over by Google and Apple. I think they have improved our lives. We have plenty of competition at the platform level (Google vs. Apple vs. Microsoft vs. Amazon). Security is better with large ecosystems. People trust Apple and Google, so they are incented to regulate themselves in order to keep that trust. Meta, not so much. The technology itself will force it to evolve to stay competitive (see the good comment on ChatGPT here and think about how ad blockers will force Google to innovate to find new revenue streams).