Location information leaks from ad targeting. Maybe that doesn't bother you, but it might bother other people.
One way it leaks is that an advertiser generally knows what ad campaign a user clicked on. So if that ad campaign was only sent to a particular location, you know where the user was when they clicked on the ad.
Ad location targeting at best is in a 3 sq. km area, not 1 km. For an average city in the U.S. (like St. Louis) that's an aggregate density of 6,000 people.
Knowing that someone in 6,000 people in an area clicked on your coffee ad is not leaking GPS data. Which is probably why bounty hunters are buying the data directly from cell carriers and not running targeted Google ads.
It's a leak of location information, which is what this sub-thread is about. The location information was only inside of Google, and now the advertiser has it.
It's also not "one in 6,000 people" if it's someone who signs up for some service via the ad. That particular someone might not realize that the service knows their location.
And that's the same deal with Geo IP, which this sub thread is about.
The point in 1:6,000 is that whether it's ad targeting or just regular internet traffic, there isnt enough specificity to "leak" data about the person.
You can split hairs over, "Well what of they join that 1:6000 with other PII" and it's the same with Geo IP. The user gives up that data but it's not different than just visit the page normally.
Website owners already know your rough geographic location via IP address. If it's that big of a concern for you, you should route 100% of your traffic through a VPN.
One way it leaks is that an advertiser generally knows what ad campaign a user clicked on. So if that ad campaign was only sent to a particular location, you know where the user was when they clicked on the ad.