Ad networks naturally form monopolies. Ad matching algorithms can make better matches (more revenue per click and more desirable product for advertisers) the more ad campaigns they have to select from and the more display properties they have to publish ads to. It's a market extremely prone to winner-takes-all.
I say this as someone who worked internally on an ad system which used to be one of Google's primary competitors. I specifically worked on projects intended to address these runaway network effects. It wasn't possible and the company exited the business (they actually tried to be a customer of google's because it'd be more profitable to just show their ads, but this was blocked on antitrust grounds ...)
The monopolistic aspects of these entities aren't necessarily the customer facing products they market to the masses.
Just because something has a network effect doesn't make it a natural monopoly. The network effect means the network may need a minimum size in order to be viable, but that doesn't mean there can't be a dozen competing networks of that size that are all viable.
Why would you spend more money to buy a less-well performing ad?
Why would you display a less well paying ad on your web property?
If you're doing search ads, you cannot fund the megawatts of infrastructure to keep pace with google if you're working with smaller margins. This is not theoretical - adwords/search has destroyed almost all of the competition, save those who have other ways of funding themselves.
So the smaller networks would make less money than Google. Less money than Google isn't the same as less money than is necessary to run servers.
And servers get cheaper as time goes on. The number of "megawatts" you need to serve a given number of ads is exponentially decreasing over time.
> This is not theoretical - adwords/search has destroyed almost all of the competition, save those who have other ways of funding themselves.
You can't ignore the companies they bought out when you say the others went out of business. They don't have any reason to buy a competitor that was just going to fail anyway, and they empirically have bought several competitors.
This is why search engines have consolidated into Google and Bing. It is not currently possible to create a search engine.
> The number of "megawatts" you need to serve a given number of ads is exponentially decreasing over time.
This is completely backwards. This is not theoretical, this is a fact. Please understand that I spent half a decade building bigger and bigger datacenters because the rate of infrastructure expansion to keep pace with growth is absolutely massive.
> You can't ignore the companies they bought out when you say the others went out of business.
They did not buy out their primary competitor. I have an insider view here and while I can't share the details I can say that you really have no idea what you're talking about.
Figuring out which products have high margins and entering those markets is an increase in competition.
Promoting their own listings is only a problem if the competitors can't reach those customers in any other reasonable way, which isn't the case because there are many other online retailers and it's quite feasible for a medium sized competitor to sell directly to the customer on their own website.
And how would you stop them from simply replacing their competitors? From the article "Budding companies that resist being acquired are pruned through flagrant copying." Would you pass a law that said Facebook can't expand their business to do anything a competitor is doing? I can't see how that would fly.
And the challenger has an advantage against a conglomerated incumbent because they can add features that reduce profits in ancillary markets the challenger isn't in. The incumbent won't want to copy that and cannibalize their other market, but the users want that feature so the challenger gains market share.
Competent competitors will do that naturally. See TikTok vs Reels.
None of these things are a natural monopoly.
Facebook only stays relevant by buying its competitors. You don't even need to break them up, just disable acquisitions. Then Instagram would be separate, WhatsApp would be separate. You don't even have to break them apart, just wait for the next one and don't let them buy it.
Google is the biggest search engine but search isn't a natural monopoly. Their dominant position comes from all the vertical integration. Google owns YouTube and YouTube is featured prominently in Google Search results. To get Android/Google Play, OEMs have to use Google as their default search engine. To fix this you don't have to regulate the search results, just separate the search engine from the other companies, and stop allowing them to buy competing ad networks.
Amazon doesn't really even have a dominant market position. They're not a railroad, they're Internet Walmart. They're just the biggest player in a highly competitive market. They have no capacity to significantly raise prices or people would immediately switch to any of their hundreds of competitors. Most of the complaints are from competitors butthurt that Amazon is keeping everyone's margins slim -- but that's what they're supposed to be doing.
A railroad is a natural monopoly. You can't really break up a natural monopoly, so you have to regulate it, which is terrible.
None of this tech stuff is like that. The problems here come from consolidation through mergers and vertical integration. They're unnatural monopolies with clear lines across which they can be broken up without long-term regulation. Just stop letting them buy their competitors.