I’ve never considered Alexa to be that accurate in its rankings, due to its flawed methodology:  “Alexa’s traffic rankings are based on the usage patterns of Alexa Toolbar users and data collected from other, diverse sources over a rolling 3 month period”.  What this means is that Alexa’s traffic estimates and rankings are primarily determined by users that have installed their toolbar.  Toolbars are a dime a dozen, many are considered spyware, and the only one worth a damn and that I would actually install is the Google toolbar.  While many thousands have probably installed the Alexa toolbar, it’s hardly a way to measure anything since it represents such a small sample of traffic data, and is not a good indicator of typical internet browsing behavior.

Two alternatives to Alexa are Quantcast and Compete.  I’ve found that Compete seems to have a much more accurate estimate than Alexa and Quantcast, plus it allows you to compare your site/blog with competition.  The compare feature is both a blessing and a curse.

As you can see from the numbers below and Internweb.com’s actual numbers, Alexa is not accurate at all.  Alexa’s founders sold their company at the height of the internet boom to Amazon.com in 1999 for 250 million in Amazon stock.   Wow.

• Alexa has Internweb.com ranked 439,457 and doesn’t estimate uniques.
• Quantcast has Internweb.com ranked 300,351, with uniques estimated at 3,879 per month.
• Compete has Internweb.com ranked 125,735 with uniques estimated at 14,000 per month.

Since I have Google Analytics installed on all my sites and blogs (and you should, too!), the actual uniques in January 2009 for Internweb were 28,801, which is substantially more than the estimates of any these popularity checkers.

The conclusion is to use Compete for you popularity (ego) checks.