It’s interesting that different statistics gathering services record different results for the same measured quantity – in this case the answer to the question “how many people browse to Climate Audit?”.
Ignoring the last data point, which is just the cumulative total for the part of the day when I captured the picture, you can see that CA averages 9,000-12,000 during weekdays falling to 6,000-8,000 on weekends.
Next Google Analytics, which is a fairly comprehensive package of almost everything you would want to know about who is visiting your website, where they are, which OS and browser they’re using, whether they dress to the left or right etc. This service produces many reports on this and many other things. I discovered that we have a single visitor from Greenland and some rather assiduous readers on the island of Jersey.
The difference between them I can only ascribe to the caching used on the site. According to Google Analytics, the average time spent on the site is of the order of 4 minutes at a time, which is much less than the cache takes to expire. So of Google does not take account of caching and WordPress.com does, then it might explain the discrepancy.
I’ve no doubt that another method like Alexa would yield a different result again. Real world statistics gathering is not an exact science, not even for websites.