[safety1st]

I realize this is anathema to Google, but they could always pay a human being to investigate the domains that receive the most spam reports, and make a decision about it based on human judgment.

One wonders how much bullshit might have been solved through a .005% increase in Google's payroll...

[titzer]

Having worked at Google for almost 10 years, I can say that almost any solution proposed that includes "we'll have a person look at this with their eyes" would be instantly discounted as unworkable. The reality distortion field towards automatic everything with computers is intense.

[safety1st]

It makes sense for Google, the company has learned that they can focus that intensely on reducing costs and get away with it. I think it is our fault as a society (or perhaps our government's fault) that we have not created an environment where the competition is stiff and Google needs to invest more heavily in quality to stay relevant.

[janalsncm]

That is literally how their search ranking team works. Here is a search strategy, here is another. Tell me which is better based on a bunch of random queries.

For spammy websites, you don’t review every one individually. But you have an algorithm that e.g. downranks them, and you AB test it.

They probably already have this strategy now, just not deployed for one reason or another. (For example, it might be expensive and unreliable to scan every website.) But if Google loses enough traffic to a competitor, they’ll be forced to do something.

[meetingthrower]

How much review would it take? How many websites until you get to 99% of traffic?