alexa and traffic

[real]

Alexa generally shows correct traffic, even if there is a temporary bump it disappears when things are averaged the next day.

But this time there is something strage. Alexa shows 10 percent jump when compared to last week, but there is 20 percent drop in actual traffic stats and adsense stats show. So is it possible that some traffic going to domain leaks (redirected elsewhere) before it arrives at the site because of a server/dns problem (or abuse)?

[andriusk]

Alexa is just showing visits from people who has Alexa bar installed, so numbers can be wrong.

[real]

I know that much. I'd appreciate possible answers to second paragraph then: can traffic be lost via server, trojan related redirects.

[andriusk]

Sorry, but I'm not pro in that second part ... :(

[just-4-teens]

alexa is no where near accurate untill u get in top 100.

i have a site that gets 4,000 uniques with a better alexa stat then a site that gets 12,000 unique

[real]

How can you check if traffic going to a domain leaks from time to time for some ip ranges?
For example if alexa traffic is very high, but actual traffic is low; this would be a reason to suspect that traffic from links or type-ins might be redirected elsewhere.

[echelon]

Have there been any studies that show how accurate Alexa ratings are? I have a rating that I am very curious about, and I would be very eager to learn if it is accurate.

I do suppose that with only toolbar users as the sample that ratings would be slightly skewed. Is there any better way to rank websites than this?

[tstaut]

I dunno of any specific studies but it's pretty common knowledge that not only are they inaccurate, they're also easily manipulated.

I think "skewed" is too subtle of a word for how unimportant their numbers really are.

Is there a better way to rank web sites? I'd say traffic- uniques, visits and page loads. :)