How accurate is Alexa for traffic measurement?

In my case the stats all came from the same webtrends system.  But that is a big problem comparing form disparate sources.  In fact i volunteered for a similar exercise on another list and gave up because the data I was getting was patently wrong.

However if we limit to (yes under 100,000) I hope the people would be sufficiently professional to know how to measure visitor sessions per month.  (I concluded last time that this was most likely to correspond to alexa but can't remember why).

So if anyone is up for it I am happy to correlate data.  and share results of course.

Personally I would be most interested in b2b (ie corporate users) but thats just me.  If contributors indicate b2b or b2c I can do two graphs.  Not much point being finer than this I think.  Yes the profiles of b2b audiences are quite different - in fact I would expect a big skew here because the big guys are not going to allow things like alexa toolbars on their corporate PCs..

So I suggest

1. Alexa <100,000

2 send visitor sessions per month for a recent month

3. Indicate b2b b2c

4. Subject Alexa ranking data

5. to bob@textor.com

If I get enough I will graph the data and distribute.  Tell your friends.

Bob