April 16, 2008

Alexa Overhauls Ranking System

Duncan Riley

83 comments »

alexa.jpgEveryone’s favorite web statistics whipping boy Alexa has announced a major overhaul of how it compiles traffic figures.

The biggest change is Alexa’s decision to drop exclusive reliance on the Alexa toolbar for traffic data, with Alexa now aggregating data from “multiple sources” to compile its statistics and web rankings. As part of the move, historical data from Alexa is no longer available, with data now only going back 9 months (we presume calculated using the new methodology).

Alexa is spinning the decision as a step forward without admitting to previous flaws:

Your ranking wasn’t wrong before, but it was different. Alexa toolbar users’ interests and surfing habits could differ from those of the general population in a number of ways, and we described some of those possible differences on our website. While the vast majority of sites’ rankings were unaffected by such differences, we’ve worked hard on our new ranking system to adjust for situations in which they could matter. The new rankings should better reflect the interests and surfing habits of the broader population of Web users.

A search of tech blogs saw many with significant drops in rank, where as political sites have had big boosts. For example TechCrunch and the Drudge Report were tracking similar figures on Alexa prior to the change, where as now the Drudge Report is a mile out in front.

Although regularly derided in the past for its often bizarre results (like YouTube having more traffic that Google), Alexa has continued to maintain popularity due to its broad global reach and completely free service provision. Time will tell if Alexa has done enough to appease its strong and vocal critics.

  • Sphere It

Trackbacks/Pings (Trackback URL)

  1. Alexa Updates Its Ranking System | SYP
  2. WinExtra » From the Pipeline - 4.16.08
  3. Alexa Stats Gets More Accurate Results
  4. Alexa Announces ‘New’ Rankings
  5. El nuevo ranking de Alexa | Tecnodatum
  6. Alexa Announces ‘New’ Rankings | To Bloggers
  7. New Alexa Rankings | Sha Money Maker
  8. Alexa Updates » Blog of Eli Burford
  9. MAIN-Blog.de » Blog Archive » Alexa will Traffic-Messung überarbeiten
  10. Alexa, alexa rank, alexa toolbar » ALEXA RANK SİSTEMİ DEĞİŞİYOR » Elektroseyir(Beta)
  11. Alexa improve rankings system
  12. Alexa improve rankings system
  13. Neues Alexa Bewertungssystem | myPEAK.net - Ein StartUp erzählt
  14. Alexa Ratings Algorithm Changed | Crenk
  15. Internet: Alexa cambia su ranking «no estaba mal, simplemente es diferente» - ALT1040
  16. TechCrunch Japanese アーカイブ » Alexa、ランキング・システムをリニューアル
  17. Alexa Overhauls Stats: You Seeing Changes? » David Risley
  18. Alexa modifica su ranking y nos da malas noticias
  19. Exploring beyond the toolbar, Alexa overhauls rankings « IT Spot
  20. Alexa Updating Ranking System - Bargain Sites Affected? | Bargain Blog
  21. ricdes dot com
  22. HighFivez.com: Your Guide Through The Blogosphere. Find Blogs and Blog Posts.
  23. The Numbers Guy : Why Good Data Can Be Hard to Find
  24. Test » Blog Archive » Alexa cambia su ranking «no estaba mal, simplemente es diferente»
  25. Alexa updates ranking methodology - Untethered Dreams
  26. ArmorBlog » 1k and Counting

Comments

RSS feed for comments on this post.

  1. R.L.

    I was noticing today on my SearchStatus bar the wild Alexa change, thanks for the heads up!

  2. bob cobb

    Wow, I went from around 150k to around 85k :shock:

  3. Trisha

    O.o this would imply that more techie people use Yahoo, and more political people use Google, if you think about it.

    Somehow that just doesn’t make sense…

  4. Jay Smooth

    “Your ranking wasn’t wrong before, but it was different.”

    Wow.. these guys are ready to run for president.

  5. Sean

    I noticed this today and just thought it was a bug. Clicky’s ranking dropped from ~7000 down to ~32000.

    Oh well, I say fuck Alexa anyways, it’s horribly inaccurate. It has shown Clicky’s traffic as basically plateaued over the last 6 months, when in fact it has more than doubled during that time.

  6. Will Johnson

    Seems like a smart move to avoid being completely plowed under by Compete.

  7. Thomas

    I still prefer Compete.com…. in fact, I bet this “change” was in hopes of catching up with Compete and Quantcast.

  8. CARversation.com

    About time, their rankings were way off before.

  9. bflora

    Mike, good catch on the political blog boost. My News + Current Affairs Magazine just jumped 70,000 spots!

  10. * Miss Universe

    The rise of companies like Compete.com probably had a lot to do with them being forced to see their strong bias towards hi profile Tech websites

  11. Ted Rheingold

    Yeah, great. So what?

    Why do they still use statistics of per million humans. Who talks likes that. What advertisers ask you your reach per million or page serves per million.

    Alexa rank means nothing, so why make that the only data point you can take away. They need to really move into the 21st Century. Or throw in the towel if glatial change is their only pace. I almost never make negative comments here, but Alexa, you deserve it.

  12. Pierre

    Let’s see…Alexa ranking system:

    1) Reach per million out of unknown no. of millions of users —-> I don’t care
    2) Rank out of unknown no. of websites ——> I don’t care
    3) Pageviews per million out of unknown no. of millions of users ——> I don’t care

    Why won’t they rank websites the way data analysis tools and compete and quantcast do, which is by monthly uniques (that actually mean something)? Alexa is on it’s way out of people’s hearts and minds, fast.

  13. Samir

    I still really like compete.com. but this is an interesting move, I’m curious about if it helps bring back users that used to use alexa and moved to other services. I doubt they’ll be able to overcome the already set mantra that “Alexa ranking means nothing”

  14. Martin Murphy

    So, what is alexa’s new alexa ranking?

  15. Dan...

    long overdue…competitors like GOOG/Urchin and IZEARanks provide data from the inside out (more accurate), rather than outside looking in (estimation/sampling errors)…so what are people seeing for their new #s vs actuals?

  16. J.P.

    A comparison between 5 sites, that I happen to know the real stats, shows that this is worst than before.

  17. Winston

    we’ll this is a good move for alexa. the results they give before is less reliable, or not at all.

  18. rodel

    now i know why.. im traffic rank went down.. tsk tsk tsk.. it would surely affect ad revenue this time.

  19. Scabr

    New Alexa ranking pulls down Twitter from about 500 to 1,970

  20. Pavel

    Even the fact our ranks has dropped from 5k to 8k does not make me mad when I see that Alexa data is now much more accurate then before.
    Finally the countries traffic numbers are way to closer to the real information we have from Google Analytics.

    And even better fact is I see more realistic “side-to-side” stat numbers for our competitors. I’ve checked several of them who has shared with me their real Google Analytics numbers and now Alexa gives me pretty close (still not 100% of course) details as well.

  21. Search The Tail

    South Korea is now in the top 4 countries that the most users of our search engine are from! It ranks 12K. Other tools tell us it is in the bottom 100.

  22. YouYap.com

    My website ranking went up. Not sure how it happen but looks good.
    YouYap.com

  23. dhiraj rawat

    my website ranking has improved but it is not showing graph…http://www.trademart.in/

  24. gaurav

    my first holiday my holiday guide and package site has good traffic but it is not showing traffic graph why? can u tell me plzzzzzzzzz…..

  25. Search The Tail

    @gaurav & dhiraj rawat
    Graph is available only for sites ranked in the Top 100K.

  26. James Schramko

    A great improvement for data. I guess people like compete.com will be pushing alexa to be more accurate. We can now track non-internet marketing markets better.

  27. Mystery CEO

    Interesting - I’m assuming they’re not revealing where that other data is coming from ;)

  28. David J.

    Lots of folks in the comments are talking about Compete, but check out Quantcast. They have the superior solution — tracking stats at the source… the publisher’s site… and the site makes available as much or as little as they want to, but it’s 100% sampling (which is not sampling at all, of course) and their solve really can’t be gamed to the high side.

    Doesn’t Compete only show US stats? Only 21% of the Internet is in the US (according to comScore), so someone with a global solve that is also the superior solve will win. People are figuring out that that is Quantcast.

    This craziness from Alexa helps to marginalize Alexa data and embolden the cry for a real solution that is also free — Quantcast!

    And, no, I don’t work for Quantcast, I just have checked out all the options very carefully as one of the largest sites in the world…

  29. Emily

    Why is everyone such a fan of Compete? My experience is that Alexa reflects the traffic trends on my site much more accurately. Compete has never seemed to correlate particularly well, while Alexa has generally risen/fallen along with google analytics. The relative height/depth of the of the peak hasn’t always been right on, but over time it has definitely been the most accurate of these services, as well as the most timely.

    I also don’t see any value in showing unique visitor numbers that have almost no bearing on reality: Compete monthly uniques for our site have generally been much smaller than our DAILY average, and seem to be similarly ridiculous for other sites in our category.

  30. Life on Mars

    Just find alexa has changed a lot, Dosh!

  31. Jon de la Vega

    My site got yesterday to the top 1000 (It had a 936 rank) and today is down to 13273. It’s funny because yesterday I added the site to the Rails Top 100 list ( http://rails100.pbwiki.com/ ) and it was supossed to be the 3rd page in alexa coded with ruby on rails.

  32. Sudoku Maniac

    I am tracking my Online Sudoku Solver site ( http://www.sudoku-solver.net/ ) with Alexa. It says my pageviews have gone up by 56% in the past 3 mos ( I guess it means 3 months). Actually my page views have gone up by more than 500% in the last 3 months.

  33. A. Naser

    Alexa now aggregating data from “multiple sources”

    Anyone have an idea what are these “multiple sources”?

  34. AbdulBasit Makrani

    I woke up today morning and normally checked Alexa traffic ranking which I usually checks every morning and was shocked to see that from 49k it’s now changed to 26k!

    I am very happy with the recent change in their algorithm

  35. heddy

    Alexa is great for *rough* overall worldwide comparisons of sites. But they don’t tell you their sample size (# of toolbar users, country distribution, etc) and they don’t tell you their new ’sources’.

    Compete looks nice but only tracks US users from a small number of ISPs, and won’t give you an idea of their sample size.

    Quantcase do the same as Compete for the vast majority of sites but some sites (’quantified’) embed Google Analytics-style tracking pixels in each page and provide all the info. Such as:
    http://quantcast.com/answers.com
    When they do that, it’s all good. But for most other sites, it’s just the same as Compete.

  36. sean

    i haven’t used alexa for quite a long time…

  37. Stefan

    First step in the right direction - better late, than never!

    However, still a lot of question go unanswered, e.g.:

    - What are these new sources?
    - Will analysis of non-.com-Sites improve?
    - Or will it still be focussed on the US?

    Folks: Ever thought about involving the community out here when it comes to improving the site?

  38. duvet-dayz

    This look much like work in progress.

    We checked a few web sites and besides that their ranking has dropped substantially there are now also whole countries missing (e.g. USA) with at least one site we checked.

    In the past we have, for example, seen that some topics (odd / humor) that went up high or to the front page of Digg.com / Del.icio.us etc were massively over-counted while for example more technical topics that went to the front page of social bookmarking sites - while the same traffic occurred - were only counted around 10% of their actual size. I believe that says something on the composition of their sample. And some leading Desktop security products identify their applet as spyware anyway.

    We have audited / externally measured traffic numbers taken directly at our servers and those have continuously differed from Alexa like night and day.

    Having been so inaccurate in the past we have just stopped to take any notice of Alexa’s numbers. But we still see them handed around with some advertising (e.g. AdBrite) or other services but believe this is mainly because of the availability as a web service.

    Given that since a year or two others provide much more accurate traffic and usage figures to a broader audience, Alexa must have been loosing massive ground with their client base lately. That might have triggered the change towards taking more appropriate approaches to measure traffic to web sites.

    Maybe Alexa will improve now but it will take a long time to overcome their reputation destroyed in the past.

  39. jeff

    yes nice to see …my Blog has moved from 116,000 to 87000..kudos to Alaxa!!

  40. Wes Mahler

    We’re up, we digg it :)

  41. Max

    @heddy - Just to clarify, Compete actually aggregates data from a number of different sources, not just ISPs. We’re also pretty upfront about the size of our panel (2 million US consumers).

    Glad to see sites ours forcing Alexa to improve. I just wish they we’re being so black box about the “multiple sources” they are adding in.

  42. beopenordie

    can everyone please go get quantified , if your not then it looks like your either not in the loop or have something to hide. compete.com is a hoax and a joke.

    get quantified! quantcast.com

  43. Vero Pepperrell

    Have people ever taken Alexa seriously, and what benefit has it ever had other than the “mine’s bigger than yours” comparison?

    *Shakes head*

  44. Jim

    I agree that Compete is useless, Alexa somewhat better, but only Quantcast is really good (for sites that use their tracker). I think the future of this is Quantcast or another tracking-pixel based service.

  45. Mark

    my site was pushed bak from 4.9K to 17K :( so sad

  46. Carl

    Funnily enough in reference to poster #44, I found quantcast to be more than useless, actually wildly inaccurate, alexa a little better, and Compete much better.

    This Alexa change took one of my sites from ranking c. 250,000 to c. 75,000 today. Quite a difference!

  47. Marc

    well its about time they got with the times

    compete.com uses multiple sources to generate numbers, so just by that fact alone they are better than alexa.

    in addition, compete.com provides that information in useful ways, like reports, much better charting and browser addons.

    i could actually go into way further detail….

  48. Josh

    It is best that Alexa is changing its ranking system. It is about time! The system was easily manipulated before and was not precise. At least now webmasters can look to Alexa for reference on traffic.

    Josh
    http://www.abigarcade.com

  49. Josh

    It is best that Alexa is changing its ranking system. It is about time! The system was easily manipulated before and was not precise. At least now webmasters can look to Alexa for reference on traffic.

  50. Trond Erik Jacobsen

    Hmm…

    This probaly make my little project worse to complete, as I need to include alot more sources for marketing the site to get in the top rankings.
    My idea was to get a website amoung the top 100 alexa ranked websites within 365 days. http://www.alexa100in365days.com. I`ve had a website rank 800 in alexa before, so I had a good feeling about the project before Alexa changed their system.

    Trond Erik
    http://www.Alexa100in365days.com

  51. Sportsrapp

    Sports websites saw huge jumps in alexa rank. I figure if you do good work and pull decent traffic, your numbers should be reflected fairly on which ever service you prefer.

  52. Ashish

    This is a good news as they will now consider more sources of traffic and not only reports depending on alexa toolbar.

    Traffic details will be now more accurate and real.

  53. Dennis

    Heard that Alexa added results from dmoz.org to their algorithm.

  54. alexa widgets redirection

    I heared the announcement that alexa has changed its ranking algorithm. I am little happy with this change because my website got little improvment. see new alexa ranking for http://www.alexa.com/data/deta.....ehotels.in Fortune Park Hotels Ltd.

  55. Gireesh Sharma

    Even my website’s ranking dropped, though by couple of thousands only.

    http://www.empxtrack.com

    Strangely, we are seeing a consistent increase in the number of unique visitors through Statcounter.com, the changes in Alexa ranking are slightly dis-aligned.

    Wondering, how to make our visitors install Alexa Tool bar!!!

    Gireesh

  56. ASIF IQBAL

    Change is not easily welcome but it’s always for good!

    We should talk about which are other sources where Alexa is taking data.

    Is it Ranking.com?
    Is it Dmoz.org?
    Is it Google’s PR?
    Is it StatBrain.com?

    Plz share your thoughts here

    ASIF IQBAL

  57. alexa ranking widgets

    I heared the announcement that alexa has changed its ranking algorithm. I am little happy with this change because my website got little improvment. see new alexa ranking for http://www.alexa.com/data/deta.....ehotels.in Fortune Park Hotels Ltd.