Mon 17 Dec 2007
I really don’t understand blog statistics - and the more confusing it gets, the more I become fixated by them!
No two sites seem to measure them the same way - or if they do then they end up with different answers in any case. Why is it so difficult to measure how many people actually visit a certain webpage within a given timeframe?
It’s not exactly imperative that I know, and I didn’t start this blog in the hope that millions of people would hang on my every word but it’s nice to know you aren’t here just talking to yourself!
Somebody, somewhere please put down some guidelines on what constitutes a real bona fide hit to a webpage so we can find a stats method which allows a level playing field for everyone.
It would certainly cut down on a lot of confusion for people like me!
December 17th, 2007 at 5:43 pm
You’re right, each metrics program works ever so slightly different. Some count the robots and spiders that weave their way through you site when they are indexing your pages for search engines, others don’t. Some use JavaScript technology (to avoid counting the spiders, I think), some ignore this. Unique hits can be defined by each user in some cases, some saying 1 minute, some saying 24 hours.
I would try a few programs and take a reasonable guess between all the answers. Because ultimately, that is probably the best you are ever going to achieve. A reasonable guess.
Whenever I look at stats I tend to look at trends more than the actual numbers. Am I going up, or down? Which day is busier? Which month is slow? What caused the spike on Saturday? What kind of post is generally more popular…
…looking too closely at each individual number will lead, eventually, to insanity.
*Just looked at yesterdays numbers for my site, comparing my host to Google Analytics. The difference between the two was ~6k, Google being the lower and more accurate representation of human visits to the site.
December 17th, 2007 at 5:45 pm
Forgot to finish that paragraph off properly:
And then pick the program that is closest to your estimate from all the choices.
December 17th, 2007 at 11:34 pm
Ollie »
It’s just so frustrating - and my stats aren’t quite so high that there could be a 6,000 difference between two stat sites!
I did mean to look at the Mint package you said you used, but never got round to giving it a try - might have to give it a go if it’s any good…
December 18th, 2007 at 1:25 pm
I love Mint and I’ve been advocating it for well over a year now, but even Mint has its flaws.
For a start, it costs. Only $20-30 (one time fee, occasional update fees every other year or so). But that is a helluva lot more than free. However, the product is great. It looks simply gorgeous, gives (in my opinion) the most realistic view of my site in terms of numbers and is a breeze to navigate and understand. And like WordPress, developers offer Peppers, which are like what Plugins are to WP, thus extended it even further.
It is hosted on your own server though, so check your T&Cs before installing. Some hosts don’t like server-side metric programs because they can increase the loads on the servers.
December 18th, 2007 at 1:39 pm
I may look into it - $20 isn’t quite so bad if you feel you can trust what it’s telling you!
There seem to be a few things which have to be in place for the Mint stuff to work so will have to look into that and see what happens.
Thanks as ever for popping by and offering advice