Alexa Totally Unreliable?
Thursday, February 16th, 2006 by MR
Many people use Alexa as a generic “site comparison” service, but is it accurate? Here are some oddities I found:
The Blog Herald was recently sold to a private buyer, however in the SitePoint auction it was listed as having 750,000 pageviews per month, with an Alexa ranking of 26,607 (now it’s 21,586 at the time of posting.) Okay, well let’s compare that to the Alexa ranking for 9rules.com, a site that receives only about half as many pageviews as The Blog Herald supposedly receives which is about 300-400k pageviews per month. Well Alexa tells us that 9rules’ rank is 21,772 which is less than 200 ranks lower than The Blog Herald, a site that is said to receive more than double our traffic! Well either the traffic of TBH is grossly overestimated, or Alexa is messed up. Here’s a graph comparing the last few months of traffic on Blog Herald vs. 9rules:
9rules is in blue, TBH in red, and as you can see we’re pretty much even except for those 9rules traffic spikes in December and January. So for a site that receives 3/4 a million pageviews per month, Alexa is reporting their pageviews graph much lower. I’m going to come out and say that Blog Herald’s traffic is probably much higher than 9rules.com’s (to be expected, at least in our fledgling company’s first few months) so it seems that Alexa is a bit off here.
Now as a comparison, I’m going to put Business Logs up against Paul’s extremely popular Damn I’m Cute celebrity/gossip weblog. I’m staring at Paul’s stats right now via StatCounter, and I know for sure that he receives between 25-40k pageviews per day, depending on who links to him or what pictures he makes public. Now compare that to Business Logs’ daily traffic, which is extremely low — less than 3k pageviews per day. 25,000 vs. 3000, it should be an easy battle:
Alexa is off again! I know for a fact that Paul’s weblog does nearly 10x the traffic that Business Logs does, however in the graph it’s nearly even or Business Logs is ahead for most of it. That makes no sense to me.
So is Alexa a good way to measure sites’ traffic relative to other websites, probably not. But that won’t keep everybody from using it :)
Reader Comments
15 Responses to “Alexa Totally Unreliable?”
agreed - it’s quite bizarre.
When I compare my ProBlogger.net site with my LivingRoom.org.au domain here’s the ranking:
PB: 6,812
LR: 14,078
Of course LR gets 4 times the traffic that PB gets.
From what I know of Alexa it only measures those who use it’s toolbar, I guess some topics of sites get a lot more people using the toolbar than others….
Not the best measure - but with it ranking my blog as 6812th highest ranked site on the web I’m not complaining :-)
February 16th, 2006 at 3:32 am
Darren is right, Alexa only knows what folks with their toolbar tell them, not actually what is happening. Is a pretty good article explaing the problems with Alexa. As far as I know (albeit, I am often wrong, and not very well read) there isn’t a toolbar option for FireFox either.
So, if your visitors are mainly IE visitors, using the default Alexa tools your site will do better than one, say mine, that is nearly 50% FireFox users without the option of Alexa. Of course, my site does very little traffic to begin with, so probably not the best analogy;)
February 16th, 2006 at 6:25 am
I’m constantly griping about this whenever I see people using Alexa data.
Everyone has to remember that they pull these numbers based on Alexa toolbar users.
So what the data really says is that you’re doing better with Alexa users than the Blog Herald.
February 16th, 2006 at 9:15 am
Only tracking the browsing habits of those that have the Alexa toolbars != reliable.
February 16th, 2006 at 9:24 pm
I’ve always wondered how Alexa got its data, now I know. I’m sure it’d be a great comparison for those sites that appealed to Alexa and non-Alexa users alike (like search engines and news sites) but it could never accurately gauge sites with more specific demographics.
Having said that, if this is the best method of site traffic comparison we have, then it’s going to be around for a long time. Just look at how TV shows are rated.
February 16th, 2006 at 10:05 pm
Mike, the Alexa toolbar is pretty unreliable when you compare sites that are not extremely similar to one another.
However, I think it can be useful for looking at how your own site’s performance is trending and for comparing with other very similar sites.
For example, I would expect it to be more reliable at comparing my site with other web design blogs.
However, because of the way it derives its data, the stats certainly have to be taken with a pinch of salt - as do any stats really, because of the huge number of factors that can influence the final numbers.
By the way, there’s another toolbar that does the same thing - the Consumer Input toolbar, which claims to have over 2 million users.
I don’t have it installed, but it would be interesting to see how its figures compare with Alexa.
February 17th, 2006 at 1:08 am
I often say, as I work at a company OBSESSED with statistics (soundscan, big champagne, etc), that I can take any arbitrary number from any arbitrary service and make it work for any arbitrary purpose.
Alexa is as reliable as any other number.
February 17th, 2006 at 2:47 am
In my experience, most of the people who have the Alexa toolbar installed are bloggers and webmasters who want to use Alexa to track stats. Many of them only installed it to boost their own stats every time they visit their own site. Hardly anybody else even knows about Alexa.
So the only meaningful thing comparisons of Alexa numbers really tell you is whether Site A receives more traffic from webmasters than Site B. That certainly explains why Alexa thinks Problogger.net is more popular than LivingRoom.org.au.
February 17th, 2006 at 6:05 am
If anyone is interested, there is a simple tool out there that allows you to generate an Alexa chart with up to 5 sites, change the parameters, etc. Enjoy!
http://www.evancony.com/alexa
February 21st, 2006 at 2:16 pm
I can’t stand Alexa, I’ve noticed inaccuracies years ago when I’d compare my site to a near competitor. We’d get nearly the same amount of traffic, but their rank was significantly lower than our’s. As a result, companies didn’t want to work with us; advertisers, PR people, etc. relied on Alexa to find out which site gets the best exposure for their hardware reviews.) Even when I showed them my stat logs, they’d still huff and puff about our Alexa rankings.
Having a site that has a demographic consisting of tech-intelligent visitors, they usually knew how to get rid of Alexa right when they format and install Windows. Alexa is a spyware, and so they are removed. If your site has that kind of demographic, you can kiss your Alexa ranking good-bye.
March 9th, 2006 at 10:31 am
As others have said, it totally depends on the demographics of the people willing to install the alexa toolbar. When I promoted the alexa toolbar on my website and encouraged people to install it, I saw my alexa rank jump up about 20%. Just having a few dozen people with the toolbar who are regulars to my site really made my stats go up.
I have since realized that alexa rank is meaningless and no longer promote the toolbar.
March 9th, 2006 at 12:51 pm
My largest competitor used to tout his Alexa #s all the time and say he had the most trafficed site on the net.
Of course, he also had ads for Alexa on his site to try to get his users to DL the toolbar.
March 9th, 2006 at 6:28 pm
I would have to agree. Alexa is very inaccurate and very unresponsive to correct inaccuarcies. For example they have our site tied to another site that has nothing to do what-so-ever with our site. All the stats are incorrect. Very strange, we have made numerous attempts to get this corrected and nothing is done. Can’t even get a live person to speak with you??? I have to believe that they just don’t care…
March 15th, 2006 at 8:39 pm
The usefulness of alexa stats is depended on the number of alexa tool bars in use. Can someone tell me the total number of alexa tool bars installed worldwide?
May 22nd, 2006 at 8:00 am
I got here asking Google that exact question - how many people use Alexa toolbar? If it’s a big enough number, then statistically it should work, as it would cover a wide enough demographic range to provide pretty accurate results.
Couldnt find the answer, by the way, only that most SEO boys and girls are complaining about inaccuracies.
My thoughts are that since general awareness to spyware has increased dramatically (at least judging from the amount of times I see it mentioned everywhere lately), and having that most spyware removal tools recognize Alexa as spyware, usage of the toolbar has decreased in equally dramatic numbers and is now being used mostly by those who need it for their work (some webmasters, I guess - with Google Analytics I dont see why anyone would need Alexa…), and people who have no idea and just click “OK” and “NEXT” on whatever they see to get rid of it, a group of people shrinking in numbers as time goes by, and who’s operating systems dont usually last very long…
Just speculation.
I figured, while writing this, that a good way to find out Alexa’s popularity is checking Alexa’s Alexa rank.
The results are… as follows:
Alexa.com
Avg. Review for alexa.com:
Traffic Rank for : No Data
Hiding something, are we?
Another thing:
Where do people go on alexa.com?
alexa.com - 66%
info.alexa.com - 16%
download.alexa.com - 10%
redirect.alexa.com - 5%
pages.alexa.com - 2%
cgi.alexa.com - 1%
Without a number though it’s not much help. That’s it, anyway. They wont give out that information so easily, but that’s the key issue - how many people install that stupid toolbar?
Personally, although I check my Alexa rank and compare it with my competition, I rely much more on my Google PageRank, my SERP rank, and above all - my revenue.
June 23rd, 2006 at 1:19 pm
Add Your Comment
Comments are moderated because spam's not tolerated.
