Google Analytics and the story of inconsistent statistics

Google Analytics once had a proud history of giving real power to those who used it. Its incredible set of tools would in an instant let you know who was connecting to your website, how they got there and what they were doing on your website.

But in the last four years Google Analytics has slowly lost the power it once had. Prior to 2011 if you visited a website with Google Analytics installed on it, detailed information on your visit was generally recorded. Of course for those running machines with JavaScript blocking abilities, they always had the ability to block the majority of Googles data collecting abilities, but this was small in number.

Since 2011 you the web surfer have had the right to prevent Google from collecting information about you and although there are many reports which state this might not be the case, anyone using Google Analytics will have noticed especially within the search phrase area of an Analytics report a large chunk of visitor data where the connecting key phrases are ‘not provided’.

As a small business owner, knowing the search phrases someone uses to find you can be a real gauge of not only how successful the website is, but can also show what you’re doing right and more importantly what you’re doing wrong.

Recently I looked at visible search information available on Google Analytics and on multiple domains and for all sites analysed at over 90% of this traffic is now ‘invisible’ as far as search phrase analysis is concerned.

It’s become such an issue that people are finding more and more inventive ways of extracting meaning from the 90% lump of undocumented data.

Sure you can still see where people have gone on your site, but unless these are well optimised landing pages, there is little chance of understanding what key phrases the traffic used to get there.

The other issues with Google Analytics is simply it need to load to work, so this code overhead can cause webpages to load slower and on a rough day (with issues on your server) this can actually have a real impact that may not only affect your visitors experience of the website, but if Google is looking at loading speed as a ranking factor, this can affect your position in Googles index as well.

But it’s not all bad news…

If you are only after fairly general information (landing pages viewed and search phrases people used to get there) then generally you can dispense with Google Analytics all together.

The first tool in your non-GA arsenal is Googles Webmaster Tools, although again its power has been tamed by Google in recent times, without the need of a tracking code, this will give you some information about the amount of traffic Google sends to your website and the search phrases used to get there.

Although only a subset of data is now shown, it’s useful in regards to looking at comparative data. (i.e. these are the most common phrases surfers used to visit your website), it also has the ability to tie this to landing pages which between the two can give a great deal of information.

But using Webmaster tools by itself is not enough, in any situation where you refer to only one set of data you are likely to be in trouble.

If you have a website, the server it sits on will be recording a log of all website traffic. Although this log is not as rich as Analytics, it too can be manipulated to provide Key Phrase and landing page traffic data.

In fact most hosts will provide a Log analysis programs to view this information.

One of the most reliable is AWSTATS, this free application will analysis your server log and spit out reports detailing such things as amount of robot traffic, geographic location of visitors, browsers used as well as the vital landing page hits, exit page hits and search phrase analysis.

The real benefit of something like AWSTATS is because it sits on the server, it does not impact on your websites performance and visitors do not have the ability to opt out of given statistic information.

There is also not a cookies issue, as the way the data is collected does not rely on this method.

But even though log analysis programs generally provide a more complete picture of visitor numbers, it should not be seen as a complete record. For instance if a visitor’s web browser has pages cached then a server version may not be served and of course it won’t then be recorded.

Other disadvantages include not without work being able to see users journey through the site or even being able to measure conversion goals, which Analytics normally excels at.

So if you’re not selling online, you may find a combination of Webmaster tools and AWSTATS is all you really need to keep on top of your analysis needs. Certainly by using both can give you a more complete picture of your website statistics, than Google Analytics alone, in addition to removing Google ability to collect website data which it then uses for other marketing products such as AdWords and AdSense.

For anyone on cheap shared hosting, using techniques which help to remove server load is a great idea, but the speed benefits are usually fractional so it’s important to weigh up the entire picture before cutting your ties with Analytics completely.