Comments on: Please Stop Quoting Alexa Data

by: Tuke

Tue, 18 Mar 2008 20:37:21 +0000

Just like I mentioned earlier today about alaxa’s numbers on mattcutts.com site. Do not try to look at the numbers as is, they are meaningless. You can do trending and comparison of like sites of which both are in the top 100K sites, anything else like in your example of comparing seomoz and allrecipes in alexa is like trying to compare a freeway to a country they are different animals and thus can not be compared on alexa data, however you could draw some conclusions between two seo sites or recipe sites. But you can never mix those two because the population using alexa toolbar is skewed comparisons between sites with different subjects are meaningless.

by: a.erol

Tue, 12 Feb 2008 17:14:49 +0000

there are not too many stat analyzers, ranking tools on the net. alexa is one of them. not the best. but one of them!

by: Nvidia

Tue, 05 Feb 2008 19:35:59 +0000

Hi,

Nice post, but Quantcast doesnt seem much use either, 44778 shows as uniques in awstats for Jan, yet in Quantcast it shows: Global, 11,584, U.S. 5,263. Anway, can you tell me how you know what pageviews are in awstats? Alexa ‘rank’ is more than Quantcast also

http://www.quantcast.com/arcades247.com

by: Alexa ranking - what’s it all about? : David Airey :: Creative Design ::

Tue, 15 May 2007 00:06:32 +0000

[…] Dustin at WebConnoisseur states that Alexa data is tremendously flawed. […]

by: Kilian

Tue, 27 Mar 2007 11:08:07 +0000

Even worse is when people try to use Alexa for analysis of any non US or worse even non English website. Alexa in Japan e.g. is just a joke. No one is using it, so the data is completely non-representative.

I’ve got a blog and a Podcast (mainly in German), and I know that my podcast’s site has far more visitors by orders of magnitude than my blog. Still on Alexa the numbers are reversed.

by: Rich Tatum

Mon, 26 Mar 2007 18:24:03 +0000

Great observations, Dustin! Thanks for sharing them.

I’ve used Alexa data for competitive trend analysis but I’ve always provided the data with a couple caveats: first, the data is not representative of the general internet population, but only of a small subset: Alexa toolbar users. What that really means is anybody’s guess because I’m not aware of a sociological study of Alexa users. :: grin ::

And second, any traffic trend analysis is only as good as the aggregation of the trends themselves, and the trends say nothing about individual user behavior. Watching server logs and traffic analysis charts is about as effective at determining what users want from your site as measuring shopping behavior by analyzing the wear on the tiles in your shopping mall. It shows trends, not much more than that.

But people can’t wrap their minds around that — they see numbers, and numbers mean reality, so the data must mean something relevant, right?

Problem is, trend analysis only goes so far in telling us what we need to know: what promotion strategy has the best long-tail impact? What strategy produces the greatest conversion of visitor to loyal reader? Why did they leave the site? Did they bookmark the site, and will they return? How are the numbers impacted by corporate and ISP proxies? So on, and so on. Some of this can be gotten at with more complicated user interaction / logging, but that’s not cheap, and it could be “intrusive.”

Again, great article. Write more!

Rich
BlogRodent

by: Richard

Fri, 23 Mar 2007 15:45:29 +0000

I’ve seen similar scenarios where Alexa is absolutely false. Yours is worse than any other I’ve encountered, though. Thanks for sharing…

by: Tami

Fri, 23 Mar 2007 11:32:08 +0000

Face Time Security labs recently discovered a case of some companies who appear to be cheating Alexa live, while we watch.

Click on the graph, and see the latest:

http://texyt.com/Trojan+used+to+boost+Alexa+ranking+China+FaceTime+Security

People like Alexa because it seems to give an easy answer

by: JC

Wed, 21 Mar 2007 21:21:28 +0000

11,842% off? Wow! Love the analogies, by the way.

by: SearchLvr

Wed, 21 Mar 2007 13:59:23 +0000

Thank you for writing this! I’m sending a link to this post to my boss. He frequently uses Alexa data, but I’ve never been able to prove to him that the data is messed up.