by alexa's standards ur site isnt the best till its broken the top 100,000
I've been using the Alexa bar for the past couple of months as just another tool to help in building 'the picture.' Alexa themselves say that the results aren't reliable until you get below 100,000, and the lower the number, the more reliable it is.
This makes sense from a statistical viewpoint, in that the bigger your survey sample, the more accurate the results are.
How many Alexa toolbars are in use? I don't know either. Last I heard last year, 10 million had been downloaded. Which doesn't tell us anything about how many were installed or are in active use today. Wild guess at 500,000 to a million???
I recently used the Alexa bar while comparing a few sites for a small ad campaign. I think it was probably reasonable to discount sites rated over 1 million, and be most interested in sites under 500,000. These were not SEO-savvy sites, so it's unlikely they'd even know about Alexa or its abuse tactics, never mind implement them.
Any point in us comparing our Alexa rankings with say our July site stats for pageviews? Here are 3 small sites I have:
Site A: Alexa
802,000 July Pageviews
13,400
Site B: Alexa
382,000 July pageviews
51,000
Site C: Alexa
170,000 July pageviews
70,000
We might get a picture on what's a reasonable margin of error for Alexa figs.