Which Web Stats Do You Believe?

Most hosting “control panels” (should) provide web site owners with access to at least one statistics program to analyse their web site traffic. If you’ve got cPanel hosting, chances are you’ve got access to at least two or three different analytics programs, including Analog, Awstats and Webalizer.

Many webmasters also use third party programs like Google’s Analytics to crunch the numbers. Or if you’re using WordPress, you’re probably using their stats plugin.

The trouble is they ALL report DIFFERENT things – even at what you might think is such a basic statistic like – “How many visitors did my site get?”

Now… is that unique visitors or normal visitors? Does that include “people” or “bots” (search engine robots) or both? Do you worry about hits, requests, page views etc?

So… Which Web stats Program Should You Use?

The answer is… “it depends” on what information you want to know!

Hosted applications like Analog, Awstats and Webalizer use your web servers raw log files, where every single “hit” on your site (pages, images – ANY file) is recorded in amazing detail! They then work their magic by analyzing each line of your log file, making all sorts of assumptions (sometimes very wild and inaccurate), and then report the summary data in many different ways.

Of these, I mainly use Awstats, because it is able to separate real traffic from “bot” traffic. When looking at these figures, the main statistic I use to provide the “best guess” of the total number of visitors to my site is the “Unique visitors” column (and if you use Webalizer, the equivalent is probably the “Sites” column).

Some of my sites also use third party applications like Google Analytics and the WordPress Stats plugin. The amount of detail you can extract from these reports is quite amazing. But even these stats may also not be anywhere near as accurate as you would like because of these limitations:

• they rely on inserting a line of Javascript code into every page on your site, and will ONLY record information IF the user has enabled javascript in their web browser. While this is “usually” the default option, a small percentage of users DO turn it off, or still use text based browser software which does not use javascript at all (probably in the 5%-10% range)
• the Analytics code is often reported as being slow to load, and may not register the visit to that page if the user quickly clicks off it or goes to another page
• you MUST be certain the code IS installed on all pages (which can be hard to achieve on non-templated sites)

There are also some third party web analytics companies who can (for a fee) provide their own interpretation of your server log files. I can’t really provide any useful comment on this type of service as I have never been able to afford to use them.

Then there are the free “hit counter” image services you can insert on your pages. These also link back to a third party application to report on the data they collect. Again, because these services use either javascript or images to track your data, you can never get a complete picture of who is using your site. Images and/or javascript may be slow to load, or not load at all – so no “hit” would be recorded. While some of these services may be useful, I would prefer NOT to publicly display any “intelligence” about traffic to my sites. If you want to go down this path, use Google Analytics instead!

My Recommendation

Please do NOT try to compare supposedly “similar” figures from one program to another! There is NO way you can accurately compare that information!

Try out all the options you have available until you find the one which best suits the information you want to know. And then…

Consistently use only ONE program depending on which METRIC you are looking at.