Wednesday, January 18, 2006

Google Analytics Accuracy Bullshit Challenge

Have you played with Google Analytics?

Has the happy horseshit syndrome settled in yet?

My web site shows the following stats:

1 direct access
2 http://domain1.com
3 http://www.google.com/search
4 http://domain2.com
5 http://search.yahoo.com/search
6 http://domain3.com
7 http://domain4.com
8 http://domain5.com
9 http://domain6.com
10 http://domain7.com
However, Google Analytics shows:
1 google
2 yahoo
3 (direct)
4 msn
5 aol
6 http://ga-domain1.com
7 http://ga-domain2.com
8 ask
9 aolsearch.aol.com
10 search
Best I can determine is Google combines all Google sources such as Google.com, Google.ca, Google.co.uk, etc. which makes it look more dominant as a single source but overall makes the actual weight of the individual Google sites merged into one big ass murky pile of BULLSHIT!

Worse yet is from my actual log files my #1 domain referrer shows as #62 in analytics.

Sorry Google, you can fool some of the people some of the time, and AdWords advertisers most of the time, but here at IncrediBILL's Random Rants we call this BULLSHIT!

Have a nice day.

7 comments:

Anonymous said...

However, it seems as if it also merges all of the msn sites together.

Frankly, why does it upset you so? For the sake of ranting?

Anonymous said...

Have you ever heard of caching and transparent proxys?

Have you ever heard of the google toolbar (which tracks a user's movements around the web)

I can guarantee you that a large proportion of those direct accesses are not actual direct accesses, but accesses where the referer header has been removed.

You can be sure that Google have a lot better idea of how much traffic you're getting an where it's coming from than your sitestats package does.

IncrediBILL said...

Mike,

It doesn't upset me, it ANNOYS me that the source of the actual referral is obscured so that I can't tell exactly where they came from and I don't see any option that let's me see that breakdown.

Also, what you're missing is my #1 referral site showing in a real log file somehow is off the charts on Analytics and shows up as #62 which is major bullshit.

Besides, the analytics are all skewed to just those that have Javascript enabled in the first place which eliminates about 4% of the visitors.

To anon: the GOOGLE TOOLBAR?
Seek rehab.

Anonymous said...

My rehab doctor told me to clarify things...

Your log parser doesn't know how a user got to your site if they are using a firewall that filters out HHTP_REFERER.

Google on the other hand, can gather data from their own pool of resources which includes other Google Analytics users and the Google Toolbar.

He also said I should try and persuade you to visit him. He reckons it'll be a hard job to get you to think straight, but he reckons he can do it for a price. :-p

Anonymous said...

Great catch. Kudos.

IncrediBILL said...

If you understood more about how Google Analytics worked then maybe you wouldn't be so fucking stupid.

Javascript enabled stat tracking simply cannot compete with the raw data and 3%-5% javsascript being disabled (which I've proven via tracking using NOSCRIPT on my pages) can account for much of the discrepancy.

If you could read fucking English you would note I said that Google Analytics shows LESS total hits for Google ranking as #1 than my raw log shows for my actual top referrer, with a referrer present, not transparent at all.

There is a reason for the discrepancy but it's not transparent proxies, more likely it's javascript being disabled which is very possible based on the types of clients being referred from that particular site.

If it was a TRANSPARENT proxy then would my raw log file show a specific referrer with larger numbers than Google?

Does that make sense now?

You remind me of a scientist that draws a conclusion when all data that shows his results are contrary.

I don't mind being wrong if the facts matched your theory but I double checked after you suggest it and they don't.

Anonymous said...

Big brother's watching!!
enough said