Holes in google analytics

How many people slip through?

[wolfadeus]

A quick question: is there a role of the thumb for how many visitors you fail to see with google analytics?

I was wondering because I got quite diverging data (around 20 to 25% of total visitors) from google analytics and another tracker, and I have never looked into my original logfiles.

[kichus]

Google Analytics can't track Visitors who disabled JS¦Images¦Cookies.
They use First - party Cookies and the data can be diferent if your other tracking tool is using third-party cookies to fetch the data.

[edited by: jatar_k at 2:53 pm (utc) on Oct. 15, 2006]

[wolfadeus]

Thanks, I know that disabling javascripts will account for the main difference; I was just wondering if there is a guideline to how many visitors you should add to the google-tracking figures to get to the actual number of visitors.

For example, if there were 30 percent of all users that can't be tracked by google analytics for various reasons.

[gregbo]

A quick question: is there a role of the thumb for how many visitors you fail to see with google analytics?

Your question assumes that some tracking software can reliably produce the number of visitors at your site. If you really want to know, you have to somehow get them to provide you with that information (e.g. national IDs). Otherwise, it's a guestimation game.

[wolfadeus]

I see...but considering that most "losses" of G analytics are based on disabled java, can't you estimate how many percent of visitors you don't track with it?

For example, if it was 10 percent and I get 900 visitors tracked, I could assume an actual number of 1000 visitors.

I thought this percentage of invisible visitors can be estimated (though it will certainly depend so some extent on your "audience", ie target group).

[gregbo]

I thought this percentage of invisible visitors can be estimated (though it will certainly depend so some extent on your "audience", ie target group).

Well, you'll get lots of different opinions, but mine is that all of the cookie, java, etc. based methods of estimating users are inherently poor. So basically you have to decide if the estimates are in line with your performance goals.

[kichus]

Well.... It's a known factor that no two web analytics tool will show same results, it vary depend on the different terminology, technology they use and lot other factors...

Google Analytics is a Free service and is considered as a genuine tool to analyze the stats. How to make sure the accuracy is a question in every tool - even in Paid tools too.

The best and well-accepted method is to use first party-cookies to track the visitors and that concept is used in Paid services like webtrends ¦ Clicktracks and the same concept is there in Google Analytics.

Eventhough these are using the same technologies, the terminologies are different and thus the stats #*$!s are also. e.g. the term 'Visit' may not be the same definition in these tools, thus in the reports it reflects.

Am not sure to post any links here,,,... if the rules are allowing that.. i may able to give you some resource links to read on...

[wolfadeus]

Thanks, guys! I might have just been wrong in my assumption that there were accurate estimates regarding the percentage of "lost" visitors from G analytics.

You input is appreciated anyways! Best, W.

[motorhaven]

Google Analytics uses Javascript, not Java. Two entirely different technogies. You'd lose a ton of visitors if the browser were to pause to load the Java VM!

[CSE_Monkey]

Base don the actuals that we track, we estimate the number of lost visitors to be between 5-7%. We see a broad mix of traffic.

Hope this helps!

Cheers!

[chewy]

my data tells me it is well less than 10%.

For sites getting 10K visits / month, Google Analytics shows Java is enabled in 98.2% of the visitors.

if your host uses AW stats or Webalizer or any other free "stat" reporting, check for your self.

and be advised, all stats programs are going to report figures differently.

My approach is to use them as "relative" figures rather than "absolute" ones and watch for changes month to month and when there are site / ranking / or other changes rather than caring about a hard and fast number.