How accurate is Alexa for traffic measurement?

Mmm... There are many factors for (large amounts of) error. We're assuming, of course, that the site stats figures given are correct which we all know isn't always the case - actually, quite the opposite in many cases still.

As a starting point I think you'd have to compare only sites with a) high enough levels of traffic (Alexa < 100,000) and b) pretty similar user bases (because of the toolbar skew) and then you'd need to be pretty confident that the site's unique user / page view stats were accurate.

Ashley

On 17:35:34 14 June 2004 textor wrote:

Not very.  A long time I did a graph

http://www.textor.com/alex.gif

comparing known traffic levels with alexa ranking.  Left side is visitor sessions per month, x axis is ranking. 

At the interesting end the graph is so steep that what looks like a close fit is in fact miles out.  I would expect a closer fit at the left hand end because the sample is bigger, but it is much harder to get any decent stats. 

If anyone on the list is interested in sending me stats / alexa rankings (no need to say what site it is)  I am happy to correlate and graph.

- only alexa < 500,000 need apply
- send visitor sessions per month and alexa ranking
- subject Alexa ranking data

Bob Browning bob@textor.com
Textor Webmasters Ltd