8 replies to this topic

[sebastian]

Hi,

We recently launched a new website and installed script from statcounter. (www.statcounter.com) We were not satisfied with the extent of data displayed and decided to opt for livestats that was provided by our hosting provider. We thought we will leave statcounter also on just to cross check. Now here is the surprise part. Both softwares display different results. In the beginning livestats was consistently showing almost double the traffic as that of statcounter. But now it varies. I took a sampling from the month of july.

date******stc*******ls
1 ****** 0 ****** 10
2 ****** 3 ****** 8
3 ****** 4 ****** 6
4 ****** 3 ****** 4
5 ****** 4 ****** 12
6 ****** 7 ****** 6
7 ****** 7 ****** 3

stc =statcounter
ls - livestats
ignore the asterix. It is just for maintaining the width

Would someone know why this is happening?

Thanks

[Edited live link per Forum Guidelines - Randy]

Edited by Randy, 07 July 2004 - 08:18 AM.

[Randy]

Hey Sebastian,

Different stats programs use different methods of measuring traffic, especially when referring to unique visitors. So what you describe is pretty much the norm unfortunately.

I'm not that familiar with either of the stats programs you mentioned above, so can't offer a specific reason the numbers are different.

[Jbrookins]

The question is if they're differentiating between uniques or not. If one is and the other isn't, that would explain a lot.

Like Randy said, you're going to see this alot. They all seem to calculate things differently, annoying as that can be.

[rustybrick]

Agreed, this is one thing that has been a real pain.

When comparing stats with someone, you can never win.

Some programs are very strict when it comes to counting uniques, while others are very lenient.

[sebastian]

rustybricks, Jbrookins and Randy,

Thanks for your comments.

Here is what the stat software says

livestats
The Total Visits amount refers to the number of times people visited your web site during the selected report period. A visit represents one or more page requests from a web site user. A new visit will be recorded when the web site user opens the site in a new browser window or if they make a request after a 30-minute "time-out" period.

Statcounter
Unique Visitor - Based purely on a cookie, this is the total of the returning visitors and first time visitors - all your visitors.

So, how do you handle your analysis, if stats vary so much. Should I just choose one, ignore the other and hopefully the chosen one would be accurate on a relative basis.

[mal4mac]

The description of 'visit' by livestats is far superior to that from Statcounter. I would send Statcounter livestats description which should (i) enable them to tell you if they mean the same thing by a 'visit' (ii) shame them into improving their prose.

[torka]

Hi, Sebastian!

The way I use the data is not so much for absolute numbers (i.e. "we had 10,283 unique visitors yesterday") as for trend analysis.

In other words, I know that the precise numbers shown by the stat programs are going to be different between different programs, and for a number of technical reasons there isn't really any easy way to come up with totally accurate numbers.

The thing to do is simply to find an analysis package that meets your needs in terms of content and presentation, and stick with it.

By picking a stat package that presents me the type(s) of information I'm looking for in a format I find usable, and by analyzing the log data using that package over time, I can see the trends in my traffic. Do I get most of my visitors during the week or on weekends? At what time of day? Is my traffic overall going up, going down, or staying about the same? When that newspaper ad with the URL in it came out, did we get the expected spike in traffic? That sort of thing.

Also -- and probably more importantly -- log analysis can allow me to follow the paths visitors take through the site, seeing what pages they generally enter through (and where they came from to get there), what other pages they view while they're there, and -- very important -- what pages do they leave from. By analyzing and tweaking the exit points, perhaps I can encourage them to stay on the site longer, or take whatever action it is I want them to take (buy a product, sign up for a newsletter, whatever).

Loads more useful than straight "visitor counts", MHO.

--Torka

[sebastian]

Thanks everyone for your comments.

And Torka, I take your advice. I will stick to livestats and focus on trends rather than absolute numbers.

Closing my post.

[deblgr]

I had the same problem when I was with my former SEO firm -- their tracking software gave suspect numbers that didn't even come close to other results I was pretty sure I could count on, like referring URLs on my PPC program.

Then I switched to NetTracker and my numbers seem much more in line. They never sync exactly, but at least I can watch whats happening now.

In my case, beware of what you get "included" as part of a package, since sometimes you're just getting something that LOOKS like its actually performing a viable task.

Deb