Why Alexa Is Worthless

Alexa is a tool that many webmasters and advertisers used to gauge the size of a website. The Alexa ranking is based on a three month running average. The number shown is where a site ranks on the internet. For example, The TechZone has a ranking of 23,005. This mean the site is the 23,005th biggest on the Internet. Out of the billions of the sites on the net, being in the top 23,000 is pretty good, right? The only problem with the Alexa ranking is it is worthless.

The main problem with Alexa is its rankings are based on users who have installed the Alexa toolbar onto their browser. If you do not have the toolbar running, you do not affect the rankings of the sites you visit. The next big problem with Alexa is it is very easy to cheat the system. Just get a few friends to install the toolbar and have them surf your site everyday. It does not take many people to break into the top 100,000. You can even do it all by yourself by refreshing your site over and over again. Get a dozen friends to do it and you’re break into top 20,000 easily.

When requesting review samples for The TechZone, I have never had a hardware vendor asked me for the site’s Alexa ranking but I have heard from many smaller tech sites that have. I get the feeling that is the vendor’s way of denying a hardware request. They would look at the site’s Alexa ranking (say it’s 60,000) and then reply “Sorry we can only send review samples to sites with a 50,000 Alexa ranking or better.” At least I hope this is the true reason they use Alexa – as an excuse. I would hate to think a PR rep would base his/her decision to send a product on a site’s Alexa ranking.

The biggest users of the Alexa ranking are ad sales networks. They use the ranking to get a higher price from advertisers. For example, Text Link Ads (TLA) base part of its link pricing on a site’s Alexa ranking – the better your rank, the more money you can command. Adbrite does the same thing. If this is the case, then John Chow dot Com should command a lot of money for a link in the next few weeks. Why you ask? Take a look at its Alexa ranking.

alexarank.jpg

Today, my blog is the 1,421st biggest site on the internet, with a reach of 1,200! According to Alexa, my blog now get more traffic and has a greater reach than the biggest computer hardware review site on the Internet, AnandTech.

alexarank2.jpg

Now I realize I received a ton of visitors in the past few days, but I know for a fact that it did not came remotely close to the traffic level of AnandTech. Yet, there you have it – a traffic graph showing I blew right past them.

I guess I should use this to my advantage. Text links for sales! Get linked by a top 1000 Alexa ranked blog! Only $1,000 a month! Anyone still think the Alexa ranking is worth anything?


54 thoughts on “Why Alexa Is Worthless”

  1. Marco says:

    Its worth something but its not the be all end all of web rankings.

    The only way websites can be truly ranked if there was some way to track how many visitors actually visited every website on the web.

    Google is in a very good position to do this due to their amazingly free Google Analytics.

    I bet that is the reason why Google is giving it away for free to to keep tabs on which websites are actually getting hits… and how many hits at that and where the visitors come from…

  2. Knuckles10

    says:

    As I was reading this post I was thinking “well, actually, it is useful on things like TLA” but then you beat me to it. Ahh, well. Take advantage of text link ads while you can :)

  3. Well, I think the sudden rush of traffic was because of Digg effect and definitely people willing to read about ad networks are by far too many as compared to dedicated and addicted readers of AnandTech. And i am not surprised your blog has traffic higher than Anand Tech. I go to them 2-3 times a month or may be quarter. But i come here everytime you publish something new. Secondly community of people like me is larger.

  4. John Chow says:

    That’s just it, my traffic is no where near Anandtech’s level. The blog’s biggest day was 55,000 page views and 45,000 unique visitors. Anandtech averages 7,000,000 unique visitors and 55 million page views per month. As for community, Anandtech has a forum with nearly 150,000 members. I’m sure that is far larger than my user base.

  5. mad4

    says:

    Given that its the only free way of finding out the traffic of a website, other than finding an unpr0tected stats folder, I would say its pretty valuable.

    Sure the data is not perfect but when you compare sites in the same niche its the best indication you are going to get. Thats why TLA use it.

  6. themadsurfer

    says:

    I agree with Alexa being used by vendors when they deny a site one of their products for review. I know of a few sites that had been denied products for review by a vendor saying they only supplied items to sites with a rank of 50,000 or better but then turn around and the same vendor would provide an item to a site for review with a rank worse then 200,000.

    I would say that vendors need to stop using Alexa to determin what site they send products to. I have not only heard about this happening but I use to help out 4 review sites and have seen this happen first hand.

  7. I agree with John, Alexa is next to useless. It’s very inaccurate and misleading. There are still a few companies who base which review sites to send products to on Alexa though. With a review site myself I’ve ran into these companies and I always try to inform them how ambiguous Alexa’s stats are. Great article John :)

  8. Carl says:

    And you didn’t even touch on the fact that certain sites will have a readership that would never knowingly install spyware on your system, and that causes sites to automatically have a lower ranking. This is fine when you are comparing sites within the exact same category, but comparing, say, a tech site to a cross stitch pattern site is useless. And INTERNET BONANZA~! sites will be even more inflated, as a lot of webmasters use Alexa just to inflate their score.

    Alexa is TOTALLY useless.

  9. Thomas says:

    You’re right, data from Alexa isn’t really accurate. A few years ago I installed that toolbar for a few weeks and a bit later my site jumped to a place somewhere near 25,000 while I only had about 500-600 visitors a day back then :P

  10. Jason Armstrong

    says:

    So … what’s a better alternative? Something that reliably measures a given site’s popularity relative to the rest of the Web?

  11. Netcraft has a similar toolbar, and I prefer it to Alexa. But it definitely has the same problem. The fundamental problem of course is what metric is there that can’t be spoofed?

    I think the best one available is something like double-click (whom I hate) pioneered. The 1×1 clear gif, hosted by mywebmetrics.com (this service does not exist yet). Then you can see how many times it gets downloaded for each page, etc… I leave it as an exercise for the implementers to make it slightly more fool proof.

    But no matter what you do, it will get fooled.

  12. Brandon says:

    You’re right, I’ve been saying this for awhile now. Excellent article and I will be using it as reference to all the Alexa fans out there from now on

  13. marcel says:

    This has been the case for years.

  14. Handsome Rob says:

    I’m with Jason Armstrong, except I want to know how Mr. John Chow would rank a web site’s popularity? And once you’ve come up with a solution, how long do you think it will take webmasters to find a loophole to make themselves a few more bucks? Let us know John, now that you’ve hit the Digg homepage.

  15. Gordon Johnson

    says:

    It seems Mr. Chow doesn’t really get it.. “NOBODY” i know can make alexa’s “reach” parameter go from “83” to “1200” by trying to fool alexa. You can get that because of a digg spike, and a whole lot of other things (slashdotting, massive ad campaigns, viral campaigns) etc.etc. but saying you can cheat alexa en-masse is just foolish.

    I administer more than 1200 machines across seven different locations, and control three of the top 5,000 domains on the internet. We tried to raise our alexa ranking (reach and pageviews) by getting “ALL” desktops to install alexa and browse the sites.. we couldn’t make more than 10% of a difference, no matter what we tried…

    Alexa maybe inaccurate enough to be fooled by hordes at sub-30K traffic rank levels, but at larger traffic rank levels, it’s nearly impossible to mask your real traffic data..

    Mr. Chow should appreciate his day on Digg, and learn to appreciate the fact that “all surveys” are fundamentally flawed, and though alexa may not be “perfect”, it’s the best thing out there to gauge traffic yet… I’ve seen real data from Comscore Matrix et.al, and it doesn’t even come close in accuracy to what Alexa does.

  16. mik says:

    interesting post, john.

    alexa rankings get even more useless when you factor in the use of web2.0 technologies like AJAX.

    i read a good post about this subject earlier this year:
    alexa is web 2.0 enemy no. 1

  17. John: You get what you pay for. For Alexa you pay nothing and if you know that all you get is some idea of what the alexa toolbar people are doing then it is worth something (and even then no the last seven or 30 days, look for a trend over the last six months and it is worth something).

    So what are the alternatives:

    1) If you want to stay free I would recommend MSN Ad-Center Labs:

    http://adlab.microsoft.com/demo.aspx

    You will find lots of goodies around site statistics, yours and others, and the sample size is bigger (even if it is only all traffic on MSN, it is still much bigger than Alexa and more rounded).

    2) There are two other paid options for you. ComScore and Hitwise. There is a detailed post on why these services are very good and how to use them:

    Competitive Intelligence Analysis: Why, What & How to Choose
    http://www.kaushik.net/avinash/2006/08/competitive-intelligence-analysis-why-what-how-to-choose.html

    Competitive Intelligence Analysis: Metrics, Tips & Best Practices
    http://www.kaushik.net/avinash/2006/08/competitive-intelligence-analysis-metrics-tips-best-practices.html

    For what you are thinking of HitWise would be best, due to its sample size (but you have to pay for it :).

    Hope this helps,

    -Avinash.

  18. greg says:

    Man.. this argument has been discussed ad nauseum for years.

  19. Slape says:

    Yea I’ve been complaining about this for a while now, I didn’t try to cheat the system (so that part is new to me, I still don’t think you can do it as easily as you mentioned).

    Alexa is ranked on the basis of those with the toolbar installed, and many sites just use that to attribute how successful a site is. But the same goes with Google PR (page rank), it’s not an accurate measure of site success and its basically just another way that Companies boss and control the average or small time webmaster – Do this or we’ll demote your PR, don’t link to that or we’ll ban you.

    And even if people become aware of how powerful these companies are in terms of defining how much you make on advertisements, and how profitable your site is to advertisers, people will do nothing to change it. As long as we keep using these stupid rel=”nofollow” tags – we’re giving in to those companies, and letting them control how we interact with our site and users. But even those that stand to rebel against alexa and google, they still face being ostracized by the companies because they don’t use norel tags and get penalized thus for not doing so.

  20. Hypercubed says:

    How is this any different from the Nielsen ratings?

  21. You just have the shits with Alexa because you have a shit rating (Perhaps this is because no one gives a shit what you say).

  22. shoemoney says:

    actually your site is pretty on for a site about making money on the internet/seo/sem

  23. I don’t know anyone who uses Alexa and just came across this this morning:

    Also, in the vlogging world, most of use subscribe to videos via RSS 2.0 with media enclosures using and aggregator like iTunes or FireAnt, so Alexa doesn’t rank those views.

    No one can actually determine how many subscribers you have via RSS any way since all you get for stats is the number of people who recently downloaded your show over somethinkg like 24 hours.

    There isn’t a good ratings system for video blogs right now.

  24. K Visnes

    says:

    Thank you for this article. Now if only you could convince people on Wikipedia that Alexa is not to be used to see if something is notable or not. :)

  25. As a media buyer, I can safely say that we completely ignore the Alexa rankings for anything numerical.

    The trends graph can be useful, but we have access to both Hitwise and Comscore…

    In general we try and buy based on cost per page view or cost per click, meaning the amout of traffic will be verified by a third party adserver.

    We will also pull reports on the number of unique users, both from cookie and IP address methods. This allows us to know how many people saw the page views, and gives us an idea if the traffic was mainly one guy and his mates pressing refresh continually.

    In the UK, we have an organisation called the ABCe who can be brought in to verify your traffic levels. They do this by looking at the code used in your site analytics and doing some logfile analysis. This costs money and therefore only tends to happen to larger sites though.

  26. Dan says:

    just decided i’d look at installing the alexa toolbar, came up telling me firefox is not supported! so you can immediately wipe all web savvy users from their list!

  27. Jim Kukral says:

    No offense John, but this a well-known and widely discussed topic for years now. Welcome to the party. Not trying to be a jerk, just saying, there’s really nothing new to your analysis.

    I suppose it’s still good link bait.

  28. Deborah

    says:

    Jim,

    Everything you ever read is always discussed over and over. Let this discussion without YOU! Get a life.

  29. I’ve been having a look at Alexa. My company site at http://www.forbidden.co.uk/ has around 1,000,000 hits a month, mostly from Europe, including a steady 8,000 a day from Google ads. Some weeks we are up at 100,000 ranking, others at 1,500,000.

    We host video at http://pro.forscene.net/ which has around 1,000,000 hits a day, and doesn’t appear on Alexa at all. These are hosted via an iframe. The major broadcasters who use our web-based video tools (see http://pro.forscene.net/trial/ ) cannot install software on their locked down machines.

    Having said that, Google ads (outside the US, which we do not target) did increase our ranking from 800,000 to 300,000, so there is some correlation.

  30. Pingback: MYBLOG by Ouriel
  31. Tony Ruscoe says:

    I was looking into Alexa rankings some time ago (between Sep 2004 and May 2005). They’re absolute crap IMHO!

    Just by using the Alexa IE toolbar, I managed to get a NON-EXISTENT website (i.e. a random domain which wasn’t even registered) to rank around 1,500,000. (It entered at around 3,000,000 and took 2 weeks to climb to 1,500,000.) That might not sound like much, but it’s a higher ranking than most people’s personal home pages and plenty of “serious” company websites!

    During the same period, my personal website jumped from a ranking of 1,800,000 to around 300,000 in just one month. After another two months, it was ranking at around 100,000! (Today it’s ranking around 480,725.)

    I’m not sure if it’s still so easy to sway their results, but I certainly can’t take them seriously after discovering all that!

    [ Reposted from the Google Blogoscoped forum:
    http://blog.outer-court.com/forum/24539.html ]

  32. Jason says:

    Hey John,
    way to go on the rankings! Thats awesome, how did you drive so much traffic to your site?

    Jason

  33. John Chow says:

    Jason – I wrote a few articles that got picked up by some very big news sites.

  34. ElectroGeek says:

    John, Has your website been picked up by Google News?

  35. John Chow says:

    I have no idea if Google News have picked it up.

  36. Pingback: Webomatica
  37. I have read that making money sites, like yours, tend to have a higher Alexa ranking because a much larger proportion of webmasters have Alexa installed than the average web surfer. I enjoyed reading your post and have bookmarked your blog. I just wrote a blog entry on the privacy implications of having the Alexa toolbar installed on my blog Internet Security & Privacy Blog.

  38. Alexa IS totally worthless. My site had an increase of visits from unique IP addresses by 10 percent for August, 10 percent for September and now by over 30 percent for October which and the month isn’t finished yet. This will put me at a 50 percent increase for the three months average but Alexa shows the exact opposite! They show a 50 percent decrease. The fact that my site would have an increased number of visitors is witnessed by the increased interest in polygamy and marriage due to the HBO polygamy series, Big Love. Nobody can deny that. Maybe some of us just hate alexa tool bars and their users amount to those who own websites, and other techies. I’ve only had an alexa toolbar a few times and always removed it within days because, like everyone else, I hate tool bars but during the time I had the tool bar my browsing was different too. I think tool bars are typically discarded as Alexa should be.

  39. Thanks for posting this. It explains a lot. Quite a few of our competitiors have great Alexa ratings yet we know how much real business they recieve. Considering how easy it is to manipulate the ratings on it, it says truckloads.

  40. Shaun Ryan says:

    I think Alexa has some value because there is a correlation between the amount of traffic a site gets and it’s Alexa rank. I’ve posted some data backing this up here:
    http://blog.sli-systems.com/2006/11/alexa_vs_search_traffic.html

  41. Jeff says:

    The very existence of Alexa was always a questioning one.. I don’t hear about it often enough to know about it.

  42. HMTKSteve says:

    Alexa is not worthless if you weight it properly. One of the most sought after mailing lists is the one for those nose hair clippers you see in the back of magazines. Why? Because if you are too scared to buy nose hair clippers at a store what else can we sell you?

  43. Alexa is a fraud way of doing things. I know people use Adsense and even put their browsers on auto refresh and get below 100,000 on Alexa. Yet there site has nothing on that people would come for.

  44. Samer Bazzi says:

    I did a similiar post on this, have a look here:

    http://www.boohead.com/the-wonders-of-alexa

    Great Stuff John!

  45. Michael Zhao says:

    Glad to know…yet another informative post! Keep it coming!

Comments are closed.