Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

lol, just reminded me I had set up a google analytics on my blog a long time ago before thinking about it and that I have been complaining about the idiots putting google analystics for more than one year.

Oh gosh. How much google analytics did I put during the 6 months where I thought it was the alternative to parsing logs for stats on my server when I believed in the clown and migrated to it?

PS: I try to quit looking at stats as much as I try to quit smoking ; it is the only sensible alternative. I should not care about them.



There are good alternatives to Google Analytics that are FOSS and allow you to self-host.

Piwik is excellent and simple, and SnowPlow is very powerful if not user-friendly.

Server logs are pretty unreliable, and don't give nearly the depth of data that a true analytics service does. I think a self-hosted analytics service does very little harm, as the main danger with Google Analytics is the potential for cross-domain tracking.


Also, a very big danger with non-FOSS analytics and ad services is that they are all, in effect metadata for governments to sweep up. It's much easier to log a bunch of Google Analytics requests than to log someone's entire Internet stream; yet capturing analytics gives you a fairly exact view of exactly what they did online by tapping only a few services. The total bandwidth for capturing ALL global analytics requests is fairly small.

Officials can then claim with a straight face that it is only meta-data - they only need to log the request with all the data in it, not the response from the server, which is usually a transparent 43 byte GIF.

By using your own server, and modifying the default URL structure, a government has to track your site directly, which is unlikely unless your site warrants interest. And if everyone did this and had HTTPS, the NSA's job goes from cracking a handful of provider's HTTPS (Google, Adobe, etc) into cracking hundreds of thousands.


However, adblockers also block Piwik and other FOSS analytic solutions.

The argument is that it's trivial for sites to share and network their results, building up a profile and effectively tracking users across sites. It completely sucks because 99.99% of people with Piwik are not doing so, are not tracking, are not delivering spam, or ads and just want to know how people are using their site.


I used to be a sysadmin and have my tools to know what I wanted ;) https://github.com/jul/yahi

I care about stats as a junkie and I see this as vanity or OCD.

Best solution at my opinion is to not care and just practice.


That looks like a very nice solution, and probably covers what most people would want out of Google Analytics for a simple site. Like you said, for a personal site, stats are for curiosity and the vanity/need to know how many people came to the site. But still, that's not quite true - you might start writing more articles about a topic which got lots of traffic, for example. That's not mere vanity or OCD, that's site optimization.

For a simple site, weblogs probably give enough information to perform site optimization. I do think that there are lots of stats any website of sufficient scale needs to fine-tune operation/marketing which are NOT available via weblogs and in fact require some JS to be on the page.

There should be a way to get this limited information without infringing on privacy and without having data gaps from ad blockers.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: