Alexa computes traffic rankings by analyzing the aggregated historical traffic data from millions of Alexa Toolbar users. The main Alexa traffic rank reflects both the number of users who visit that site and the number of pages they then view. The former is termed reach and is expressed as the percentage of all internet users who visit the site. The latter, called page views, are a measure of unique daily visits, on average, by each user in the period under study. Alexa shows a daily, weekly, and three-monthly picture for any site and how these are trending over time.
So, if a site like yahoo.com has a weekly reach of 28%, this means that in an average day during that week, 28% of all global internet users (measured by Alexa) visited it. Alexa sounds good but it has a number of limitations, mainly due to the Toolbar population being unrepresentative of the internet user base as a whole. For example, the Alexa Toolbar is not supported on the AOL/Netscape and Opera browsers; Alexa users are disproportionately likely to visit sites that are featured on alexa.com, such as amazon.com and archive.org; and there is a strong relative uptake of Alexa in China and Korea. More importantly, the Alexa Toolbar is disproportionately used by webmasters with an unhealthy SEO obsession (like me!). Web directories, SEO blogs, and the like tend to have their traffic overstated. The other key point is that the reliability of Alexa’s data falls exponentially as you move down through the rankings.
The data set is not large enough to determine accurately the rankings of sites with fewer than 1,000 total monthly visitors. Alexa itself states that the data for any site sitting outside the top 100,000 is inherently unreliable, due to a lack of statistical significance. Indeed, if your site is outside the top 1,000,000, you will notice that, as soon as you start looking at your site through Alexa, your site ranking seems to tick up. This is actually you affecting the numbers! All the same, the Alexa data is the best of a bad lot, in my humble opinion. It may not be perfect, but it’s a useful guide nonetheless to your position relative to your competitors. I would install the Alexa toolbar in your browser once every three to four months for a short period to check your position, then uninstall it until the next time you want to check, so that your own browsing doesn’t skew the results too much.