Lies, Dammed Lies, Blog Statistics and Unexpected Spikes
Posted by Brian Kelly on 25 November 2008
There was some discussion a while ago on standards for usage statistics for public sector Web sites. I have always been a bit suspicious of initiatives which encourage use of simplistic metrics. There’s a real danger that rather than using such data to provide evidence and to inform policies, achieving a top ranking is regarded as the main objective itself. And such temptations can lead to organisations exploring ways of maximising their figures, even if this fails to achieve any underlying benefits to the organisation or the user community.
This struck me recently when I notice a huge peak in the usage statistics for this blog on 6th October 2008. Initially I thought the blog may have been ‘Slashdotted’, with one of my posts being cited on a popular service. But this, it seems, wasn’t the cause of the spike. I had, in fact, used the Adobe Acrobat software to create a PDF file of the blog posts. This every individual page to be accessed. And when I discovered that the PDF file was thousands of pages long rather than the 4-500 pages I had expected, I realised that individuals posts had been retrieved on multiple occasions, as they are also grouped by month and by categories.
I hadn’t expected such retrievals to be recorded, as WordPress’s statistics page states that “we don’t count your own visits to your blog“. But for some reason the Adobe Acrobat’s downloads have resulted in my statistics being artificially skewed.
So if you want to impress people with a sudden growth in the numbers of accesses to you blog, run a tool such as Adobe Acrobat over your blog. If, on the other hand, you feel this is unethical, then don’t do this with the aim of massaging your figures. And fopr the sake of completeness, on 6th October 2008 there were 1,206 visits recorded, but a more accurate figure, based on the numbers of visits the previous week, would be around 265.