Urchin Server Side Compared to Google Analytics

 

Ever since March 2004, I have been comparing analytics software built by the same company. The first entry I have on this was named Comparing Web Traffic Between Different Web Analytictical Tools which shows the huge gap in traffic when comparing Urchin 3.x to Urchin 5.x. Then the topic came up again in January 2005, with an entry named Web Analysis Tools & Consistancy. Which led me to an entry in late May of 2005 where I said, Web Analytics Needs Standards Bad.

So when Google Analytics came up, I signed up and let the data collect. Keep in mind, Google Analytics is Urchin. The differences are; (1) its a half of a version newer and (2) one is server side tracking (logs and 1st party cookies) and the other is simple JavaScript tracking (3rd party cookies). Then yesterday, I did a comparison of my trusted Urchin 5.7 package against Google Analytics for the month of December. I came up with a 20%, 19.79% to be exact, increase in traffic for the Urchin 5.7 tracking over Google Analytics. Ill post the daily increases for December in the extended entry area, you will notice it was as low as 5% and as high as 30%.

For a visual reference check out this chart, you can click on it to enlarge. The red is Urchin and the blue is Google.

google-vs-urchin-s.gif

I started a thread on this topic at our forums and Matt, md_doc explains why this may be the case. He says;

Using logs you get information that you would not get with javascript... like when someone that has javascript disabled comes to your site, or when a robot parses your site.

Can you believe that it can affect on average, 20% of the visitors of this site? Who are you that are blocking those cookies? Do you even know you are blocking them? :)

Forum discussion at Search Engine Roundtable Forums.December Shows Urchin Has Higher Visitor Counts then Google Analytics

Thu 12/1 21.63% Fri 12/2 28.11% Sat 12/3 12.20% Sun 12/4 12.87% Mon 12/5 12.33% Tue 12/6 24.29% Wed 12/7 25.66% Thu 12/8 24.66% Fri 12/9 19.76% Sat 12/10 19.67% Sun 12/11 5.95% Mon 12/12 24.51% Tue 12/13 22.20% Wed 12/14 25.83% Thu 12/15 24.56% Fri 12/16 23.32% Sat 12/17 20.24% Sun 12/18 10.42% Mon 12/19 16.84% Tue 12/20 20.50% Wed 12/21 18.96% Thu 12/22 21.03% Fri 12/23 29.67% Sat 12/24 18.34% Sun 12/25 13.66% Mon 12/26 17.03% Tue 12/27 19.73% Wed 12/28 20.93% Thu 12/29 18.68% Fri 12/30 24.90% Sat 12/31 14.94%

Previous story: Google Search Syntax Tidbit
 

Comments:

Dennis Pallett

01/03/2006 04:07 pm

The server-side version of Urchin also counts robots, and the Google version doesn't (due to JS). I think that's the main reason for the difference.

Barry Schwartz

01/03/2006 04:09 pm

It doesn't count Google as a visitor, it counts Google as a bot.

Search Engines Web

01/03/2006 07:50 pm

There are several reasons why Client Side may not show all the results - sometimes this depends on where the JavaScript Code is placed - Google suggests placing it in the HEAD - but the farther down the page it is - the less likely it may load before someone clicks to another page in the site. (also, the Javascript may not be on EVERY PAGE!!) Ironically, even with TWO trackers right NEXT of one another - you will often get varied results... Even if JavaScript is disabled - there NOSCRIPT tag or the invisible GIF does register the PageView - it just cant give the specific referrer information... If the referrer logs are studied - they can reveal what referrers are missing .... Some Client Side Stats allow for CUSTOMIZED search views - so as to see if a conjecture is in fact true. IF you are getting Search Engine referrers - as a test see if both the Search Engines and the Keywords are the same on Both Trackers... If not, this may reveal another flaw.

Jo

01/05/2006 02:44 am

Search Engines Web is a good summary of potential differences. To ensure the best comparison are you using the UTM to track visitors with Urchin v5? In the reports tab, you can select different ways of detecting visitors. The UTM is the most accurate and comparable with GA. BTW, there is a mistake in your comparison - All Urchin products and the new Google Analytics use 1st party cookies

XOONI

04/03/2006 03:06 am

If you setup a new profile and use IP + UA as the visitor tracking tracking method, you should be able to see the bot traffic, although overall traffic will be inflated.<br><br> I typically see a ~20% difference between IP + UA and UTM tracking methods in Urchin 5, so this is an odd one.

Brad Henry

09/25/2006 01:30 pm

I already have the Urchin 5 server side analytics and I have been trying to find information on how to get the Urchin 6.0 for our servers? Does anyone know how to get this? When it might released if it even will be released?

George Morris

01/14/2007 06:42 am

I fully believe Urchin to be wrong, not Google Analytics. I just posted about this <a href="http://www.imulus.com/blog/index.php/2007/01/13/google-analytics-is-under-reporting-or-is-it/">on our blog</a>

Ben harper

12/19/2007 12:21 am

I don't have any facts to back this up, but I tend to agree with the comment above. Urchin just feels wrong and inflated. For some of my sites, I may see 1100 sessions in Urchin on a day when there are only 97 visits recorded in Google Analytics. The page views seem to be more consistent, but still seem bloated in Urchin. When I look at my Google Analytics reports, I just get a sense that it's the exact number of people who viewed my site that day. Maybe I have had too much google KoolAid.

Tracy

12/01/2008 05:13 pm

I know it has been awhile since people have posted on this thing, but I thought that my research might provide a little insight. According to even Google Analytics themselves, the Urchin data is described as "more accurate." It actually says it in their description of the Urchin software. For instance, with our site - our information is stored beneath a security firewall, and since Urchin is run off of our own servers (compared to Google's server for free), the data logs are based upon the information collected without any interference. The information is inherently more accurate. The biggest question when comparing the two is the differences between interpreting the stored data. The numbers that should be the most comparative are Urchin’s “sessions” versus GA’s “visits”. Both sessions & visits represents the number of individual sessions initiated by all the visitors to the site; If a user is inactive on the site for 30 minutes or more, any future activity will be attributed to a new session. Users that return within 30 minutes is counted as part of the original session. Compared to Visitors, whom are counted during the initial session only, and any future sessions from the same user are not counted as additional visitors, only as additional visits. It’s a pretty simple distinction.

blog comments powered by Disqus