Any advice on how to reconcile Google Analytics vs. Awstats?

jdp · January 2010

We recently implemented Google Analytics. In doing a bit of comparison to the Awstats that come with the Convio admin side (Data Mgmt > Reports > Site Mgmt Reports > Web Usage Reports), which was our primary source of traffic metrics, I would like to know if anyone has any advice on how to read one or the other.

I have a grasp on Google Analytics and know there are tons of how-tos on the subject, but I am more concerned with reading the differences between GA and Awstats. It's pretty disconcerting to see just how different the numbers can be.

For example:

Awstats Number of Visits (Jan. 2010) = 9,800

GA Visits (Jan. 2010) = 12,800

Awstats Pageviews = 69,200

GA Pageviews = 45,100

Is it in how one defines the metric vs. the other? Am I looking at/comparing the wrong things? I figured there would be some differences, but this is more than I thought.

Is there something else about implementing GA I should know? I used these instructions: http://help.convio.net/site/PageServer?pagename=Admin_Enabling_Google_Analytics

Has anyone else gone through this recently?

I really appreciate any help. Thanks!

James Zetlen · January 2010

AWStats and Google Analytics use different methods of gathering data. AWStats uses server-side data, going through the server's logs of what requests have been made to it. Google Analytics uses a JavaScript codeblock, so it relies on the user's web browser to tell Google when it has viewed a page. This can result in different totals for a few reasons, including but not limited to:

Users without JavaScript won't show up in Google Analytics results (but this number of users is very small).
Users who clicked forward before a page completely loaded often won't show up in Google Analytics results.
If there are a number of broken images on your site, AWStats will view every attempt to load that image as a hit to the "page_not_found" page, skewing results.
Users who refresh a page multiple times will show up as multiple hits in AWStats, but not in Google Analytics, which considers that one page view.
Unique visitors (one person viewing multiple pages in a single session) are harder for AWStats to track. Google Analytics uses cookie data in a way that AWStats cannot.
Many Convio pages go through several server-side redirects, because of the architecture of Convio pages and in order to maintain browser security for donations. AWStats can count redirects as multiple pageviews, whereas Google Analytics correctly interprets these redirects; it will only record a page view when a user's browser is literally viewing the page!

Generally, Google Analytics and other JS-driven, client-side analytics systems are regarded as superior in accuracy and real-world data to older, server-side systems like AWStats. When in doubt, you should probably trust Google's results more. However, there are times when Google's results can be skewed as well. Here are some questions which can help you determine that.

Are you using Convio's funnels feature? If it's not configured correctly, it can result in multiple pageviews recorded to Google Analytics.
Load one of your pages and view the source code. At the bottom of the source code, there should be one Google Analytics code block. Are there multiple codeblocks? If you followed the directions in that help page, Convio should automatically set up a code block for you. But if your pagewrappers also contain cut-and-pasted Google Analytics code, you might see more than one codeblock there, which can also skew results.

c.stauffer · January 2010

Also, please remember to set up your Google Analytics account to exclude your office's public IP addresses if you want to not have internal employee visits in your stats. It takes a fairly simple regex expression to implement.

jdp · January 2010

James Zetlen:
AWStats and Google Analytics use different methods of gathering data. AWStats uses server-side data, going through the server's logs of what requests have been made to it. Google Analytics uses a JavaScript codeblock, so it relies on the user's web browser to tell Google when it has viewed a page. This can result in different totals for a few reasons, including but not limited to:

Users without JavaScript won't show up in Google Analytics results (but this number of users is very small).
Users who clicked forward before a page completely loaded often won't show up in Google Analytics results.
If there are a number of broken images on your site, AWStats will view every attempt to load that image as a hit to the "page_not_found" page, skewing results.
Users who refresh a page multiple times will show up as multiple hits in AWStats, but not in Google Analytics, which considers that one page view.
Unique visitors (one person viewing multiple pages in a single session) are harder for AWStats to track. Google Analytics uses cookie data in a way that AWStats cannot.
Many Convio pages go through several server-side redirects, because of the architecture of Convio pages and in order to maintain browser security for donations. AWStats can count redirects as multiple pageviews, whereas Google Analytics correctly interprets these redirects; it will only record a page view when a user's browser is literally viewing the page!
Generally, Google Analytics and other JS-driven, client-side analytics systems are regarded as superior in accuracy and real-world data to older, server-side systems like AWStats. When in doubt, you should probably trust Google's results more. However, there are times when Google's results can be skewed as well. Here are some questions which can help you determine that.

Are you using Convio's funnels feature? If it's not configured correctly, it can result in multiple pageviews recorded to Google Analytics.
Load one of your pages and view the source code. At the bottom of the source code, there should be one Google Analytics code block. Are there multiple codeblocks? If you followed the directions in that help page, Convio should automatically set up a code block for you. But if your pagewrappers also contain cut-and-pasted Google Analytics code, you might see more than one codeblock there, which can also skew results.

While we are discussing AWStats, can you explain what is the Ajax Helper? It is listed as one of our most-viewed pages, but it does not exist as a page.

Thanks!

Any advice on how to reconcile Google Analytics vs. Awstats?

Comments

Categories