Hacker News Reader

Comment by ipython | Hacker News Reader

The best part is where the author admits to using google analytics himself to track who visits his blog. At some point we all have to say enough is enough.

View on HN

Except that visiting a website is a known, expected information leak to the website owner, and for 70% of web users also automatically to Google.

On the other hand leaking the list of apps on your local computer, and to a third-party to top it off, is unexpected and thus more harmful.

I'd disagree with that.

Given that the Wacom utility is full of app-specific references and "customize your tablet, per app", I'd say that this is on par.

Ask random person, "Hey, do you know that when you visit John's blog, he sends your information to Google, too, not just himself?" and I guarantee you the answer is probably closer to 7% than 70%.

In theory yes. In reality no one cares what apps you use. They will do some big data analysis and thats it.

That's wishful thinking, not reality. In reality, many different actors can benefit from info about the used software on a particular users' PC. And now they know who to bribe to get it.

> Has this investor got a beta build from that new startup everyone's talking about? Their bets are always winning, I better frontrun them if that's the case.

> What apps could I exploit to get into this guy's computer?

> Wait why is my employee suddenly running tor browser after I involved her into this new secret deal? Better be careful with her, she might be talking to someone.

> Damn, our competitor's engineers are all running our app. Let's correlate the timestamps with our own backend to discover their accounts and push a special update to them.

Why do people use GA? Every HTTP server generates a log. That's always been good enough for me. Then again my site doesn't have ads, so maybe that's the answer.

Ohh you bring me back memories of 1999 when I was using WebTrends to analyze http access logs.

HTTP Log analysis is slow, and requires a lot of server side setup. Also, it will not give you navigation events in a SPA.

Using GA... just drop a line of JavaScript and you are done, with near real-time insights that are more detailed than an access log. You don’t need any server conf, or extra knowledge (not even JS knowledge: copy & paste the embed code). And Google gives you that for “free”... that’s why tons of sites doesn’t care about Http access log analysis anymore.

For desktop apps is easy too. The GA API is very simple: send the app id, event + any event data you want. Your dev team can do that with self service (no need to setup a service, no extra costs to handle data).

I'm going to go on a limb and say that tracking people who visits your blog is a little better than tracking app launches for the entire computer.

Certainly, although I would call it less bad, rather than better.

I think what escapes most people is that they think Google Analytics is a free product to help you track your visitors. Yes it does do that, but the more valuable product to other google customers is you've linked your website to a global ad network even if you don't display ads.

I've read a lot of comments recently about how Google Analytics is bad, but no one explain why. Can I ask why this is not something people want? Is it not anonymised?

It's anonymized to the website owner (who has your IP anyway), not to Google. That means even non-Chrome users are forced under Google's omniscient watch of the web.

When a company with the reach and market share of Google is involved, "anonymized" is meaningless. They have so many other channels to gather data (Android, ReCaptcha, Chrome, Google Search, etc.) that I'm sure it's trivial to de-anonymize and correlate GA data to a real person if that becomes profitable/necessary.

One reason why is that Google Analytics is not being limited to giving the web site owner traffic information, the analytics are also being used by Google to collect and correlate larger traffic patterns, as well as track individual users across multiple sites. These are things Google gets to see, but the neither the site owners nor the users have access to. What Google does exactly with this information is not entirely known outside of Google, though it's certainly used at least to improve personalized advertising, which many people feel is a privacy concern.

Google, and anyone who has access / can gain access to the data, whether by technical means (NSA) or by law.

GA gives Google surveillance over a large portion of the web. Even if the author of a webpage trusts google with their data, they shouldn't be forcing their opinion of Google on others. Trust isn't transitive!

Google receiving browsing histories for a single website is rude, but it probably isn't a serious problem for many websites (although the risk will depend on the nature of the website). In isolation, the fact that Alice read Bob's webpage isn't isn't very interesting, but Google can aggregate that data into s very accurate pattern of life[1].

> Is it not anonymised?

Not for any meaningful definition of "anonyms". At best GA will zero the low 8 bits of the IP address by request of the website. (The opinion of the person visiting the website apparently isn't worth considering) See this[2] post for a more detailed explanation of GA's perfunctory "Anonymize IP" feature.

[1] https://en.wikipedia.org/wiki/Pattern-of-life_analysis

[2] https://news.ycombinator.com/item?id=15167059

> Can I ask why this is not something people want?

I block it because the data it collects is none of Google's business. Being "anonymized" doesn't make it any better.

It literally is Google's business. That's what they do. They're an ad agency. They sell you.

I know that Google builds its business on the backs of user data. That doesn't change the fact that data regarding me or my machines is "none of their business" in the colloquial sense. That Google forcefully disagrees is what makes them a spy agency.

We've had visit counters since the 90's, so what?

Lol, comparing visit counters from the 90s to Google Analytics.

GA is used by countless websites. It's likely hooked into the adwords codebase so that they can track websites you visit even if that website does not have Adwords ads on it.

Reply on news.ycombinator.com