Why We Don’t Use the Alexa Rank

3 Comments

I love statistics.  I actually use two different systems to track users on this site – Google Analytics and Sitemeter.  Sitemeter provides instant feedback and a line by line look at activity – quickly giving some geographical demographics.  I use Google more for the long haul statistics.  I also check my Google Page Rank from time to time.

In the blogosphere, you’ll often hear people chatter about their Alexa Rank.  There are a lot of things to link about the Alexa Ranking.  It’s updated daily, and you can see much bigger movements.  Google Page Rank is a number from 1 to 10 (higher being better) and it can take years to move forward just one spot.  The Alexa ranking intends to show where your site rankings in the world rankings.  You can be the 1,000,000th most popular site one month and inside the top 100,000 the next, if you have a lot of traffic one month.  Alexa also shows some demographics about your visitors.

However, I do not use the Alexa Ranking.  Months ago, when I started playing around with it, I noticed some bizarre behavior.  My traffic would have a significant uptick one week, only to have a worse Alexa Ranking.  What was going on?  It took only a minute to realize the problem.  Unlike the other sites, Alexa doesn’t actually analyze the exact details of your traffic.

Instead, Alexa attacks the problem somewhat backwards to the way the other sites work.  It has people download a toolbar, and then tracks the activity of those users.  It then uses this data to produce the ranking for individual sites.  However, this doesn’t really show how popular your site is amongst all internet users – it shows how popular it is amongst Alexa users.  Even worse, the toolbar is not available for several browsers, completely excluding people who use them.

The very fact that the Alexa Ranking requires a download from their site introduces sampling error.  Sites that are more technical in nature will get an artificial boost simply because their users are more likely to have downloaded the toolbar.  Likewise, sites that talk about Alexa from time to time will get a boost because of increased awareness amongst the site’s visitors.  Ironically, this article will probably boost my Alexa Ranking.  This site doesn’t have a lot of tech articles, and thus I suspect doesn’t have a big tech crowd.  Thus, a small movement in the number of visitors can create a large swing in Alexa Rank – due to the fact that Alexa users make up a small portion of our visitors.

So, do we throw out the baby with the bathwater, then?  Not necessarily.  If you have a considerable amount of traffic on your site, the Alexa Ranking can provide information on relative performance from month to month.  Alexa can also provide details on the demographics of your visitors – although you might expect Alexa to report a group that skews younger and more male than you expect – once again due to the tech bias.  However, I would suggest using caution when trying to determine the overall popularity of your site.

3 Comments (+add yours?)

  1. Evan @40Tech.com
    May 03, 2010 @ 18:35:15

    I’ve always thought Alexa rankings were kind of bogus. I think many tech-oriented people don’t install the toolbar, because they stay away from any toolbar. I decided to try out the toolbar after starting my blog, and my antivirus program flagged it as malware.
    .-= Evan @40Tech.com´s last blog ..3 Free Tools to Encrypt Individual Evernote Notes =-.

    Reply

  2. kosmo
    May 03, 2010 @ 21:11:01

    I think you’re correct about many tech people being anti-toolbar. However, I’m thinking that the segment that DOES use toolbars exceeds the ratio we’d see in the general population – just because of awareness.

    Reply

  3. Steve
    May 24, 2010 @ 16:58:55

    Very nice article. More power to you, sir!

    Alexa.com should put a disclaimer on its website warning people that their data is “For Amusement / Entertainment Use Only” as Alexa rankings CAN be easily manipulated.

    Reply

Leave a Reply