←back to thread

863 points IdealeZahlen | 2 comments | | HN request time: 0.444s | source
Show context
megaman821 ◴[] No.43718617[source]
I don't think this article explains it well. Google sells ad space on behalf of the publishers and also sells the ads on behalf of the advertisers. It also runs the auction that places the ads into the ad space. See this graphic https://images.app.goo.gl/ADx5xrAnWNicgoFu7. Parts of this can definately be broken up without destroying Google.
replies(19): >>43718672 #>>43718693 #>>43718751 #>>43718794 #>>43718938 #>>43719033 #>>43719196 #>>43719219 #>>43719246 #>>43719395 #>>43719429 #>>43719463 #>>43720402 #>>43720461 #>>43720510 #>>43721628 #>>43722559 #>>43723479 #>>43724604 #
crowcroft ◴[] No.43719395[source]
When a media buyer puts $1.00 in on one side of the system, on average only $0.60 makes it to the publisher. In some cases less than $0.50 gets to them.

Advertising is an intentionally complex system so that companies can clip the ticket at multiple stages throughout the process. Google should be broken up, but the whole ad tech system needs to go into the bin if these problems are going to ever get fixed.

https://www.forbes.com/sites/augustinefou/2021/02/15/how-muc...

replies(4): >>43719494 #>>43719973 #>>43720688 #>>43724612 #
aiauthoritydev ◴[] No.43719973[source]
As someone who has worked in AdTech I would respectfully disagree. It is indeed complex but it is incredibly efficient. Also it is irrelevant of whether publisher earns 75% or 30% of the total revenue. What matters is how much they are earning compared to the next best alternative.

Some companies like Google are incredible at this. Google is not a "monopoly" in this space. In fact the world has far too many Google equivalents but absolutely no one comes close to Google in generating top dollars for publishers. I am saying this after working for 10+ years competing against Google.

replies(11): >>43720020 #>>43720098 #>>43720244 #>>43720256 #>>43720282 #>>43720308 #>>43721019 #>>43721231 #>>43721424 #>>43725204 #>>43725868 #
CPLX ◴[] No.43720098[source]
> it is irrelevant of whether publisher earns 75% or 30% of the total revenue

Of course it matters if a middleman is skimming off 70% of the revenue in a given market.

> it is incredibly efficient

On what planet is a loss of 70% of the resources to the matching process between buyers and sellers "incredibly efficient"?

> What matters is how much they are earning compared to the next best alternative.

Right, which is why it is illegal to prevent there from being a next best alternative via anti-competitive practices which is precisely was was proven in this trial after a detailed examination of the evidence.

replies(3): >>43720303 #>>43721236 #>>43721505 #
hamp95 ◴[] No.43720303[source]
> On what planet is a loss of 70% of the resources to the matching process between buyers and sellers "incredibly efficient"?

One where the market maker is taking up the cost of providing a market. E.g. Steam takes a 30% cut for providing the infrastructure required to distribute games. Some people/companies can do it for less but it is the best option for a majority of sellers.

If the market maker did not the seller would get more revenue but would also eat the cost directly instead of paying someone else to do it.

replies(3): >>43720375 #>>43720655 #>>43721389 #
crowcroft ◴[] No.43720375[source]
Are you saying that serving ads costs more than running a news site?

This also neglects the fact that the programmatic market routes billions of dollars intended to be spent on real media (ad placements on real news websites etc), to fraudulent mobile apps and websites and bot traffic.

replies(2): >>43731519 #>>43733041 #
1. bryan_w ◴[] No.43733041[source]
> Are you saying that serving ads costs more than running a news site?

Static text is much cheaper than dynamic content. News isn't anything that couldn't be served out of a .txt or sqlite CMS.

News providers seem wholly uninterested in providing anything better and actually competing with better content providers

replies(1): >>43736147 #
2. crowcroft ◴[] No.43736147[source]
Assuming the content is $0