Ad networks don't naturally form monopolies. That's why Google had to buy the other ones. Anything will form a monopoly if you let every company in the industry merge.
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.
It does, because the increased efficiency means ads are more profitable on the larger platforms. Bigger matching pools mean better conversion rates.
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.
> Bigger matching pools mean better conversion rates.
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.
No, this isn't true. The hardware necessary to run a competitive search engine increases over time on a massive scale because the size of the internet grows faster than the price of hardware can fall.
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.
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.