Stats from StatCounter, Google Analytics and IntelliTracker all differ

[MattyB]

I've recently re-branded a website I built using mediawiki for a local newspaper. Bygone Derbyshire

Previously the site used two stat recording tools to determine what was going on.

The first stat tool (intellitracker), was used by teh Newspaper to see how their figures were doing, and the second stat tool (statcounter) is being used by us to see how the site was doing, was it coping etc.

Now until recently the development company(me) and the Newspaper had never shared stats, but due to a technical error, the newspaper didn't have there own stats, so I let them view the stats of statcounter.

We then noticed some very large discrepancies between the two sets of stats for the same site, on the same day.

Stat Counter was showing more users, yet slightly lower page views.

I then added Google Analytics to the site, and even more confusion arose as this too recorded a different set of stats.

To cut a long story short, it turns out Statcounter actually counts stats when Javascript is turned off, which probably explains why the visitors numbers are different, but I've no idea why the page views differ so greatly.

Does anyone have any suggestions.

 

[ServWise]

Personally I would never rely on an external code driven system for recording stats reliably. There are a number of things that can make them very inaccurate and unreliable plus you are only ever recording hits to page that actually contain the code, not info on image hit, hits to files like pdf's or hits to anything other than web pages basically.

As you have mentioned javascript and image bugs can be blocked in a number of ways which will distort your stats. That is not possible in your server Web Log's.

If you want to have reliable stats then use something that reads the actual site log files, usualy this will be available from your hosting company.

 

[MattyB]

Good point, but you know stats recorded by the server that serves the pages also get skewed because of browser and ISP caching, so even they are completely reliable.

 

[MattyB]

Good point, but you know stats recorded by the server that serves the pages also get skewed because of browser and ISP caching, so even they are completely reliable.

That should read aren't - silly me

 

[epiphany]

Many things cause differences in stats, there are a lot of robots out there and the stat programs have varying success at working out which visitors are robots and which are real people. Also different stat programs have different interpretations of how long someone has to not visit a site before they are a unique again. I think you can set this variable in statcounter but not in analytics. There are also issues with javascript positioning, proxy servers and log files vs tag based stats. Basically no stat software is exact, there will always be significant differences.

Just stick with the software you like best and base your analysis on trends rather than exact numbers.

At the end of the day trends are the important thing

 

[downsouth]

I've also noticed that GA is not too accurate, somedays no data is shown even though sales have occured.

Or is just seems to show no logic and produce random data for the previous 24hours when you've not had a sale for a day or two

It all just doesn't seem to tie up