Gideon Marken: Web Technologist

Just in case you were not aware, Alexa is not a valid means of comparing sites - it's some form of an indicator, but should in no way be considered for solid conclusions or judgements on how popular/busy a site is.

The Alexa user base is only a sample of the Internet population, and sites with relatively low traffic will not be accurately ranked by Alexa due to the statistical limitations of the sample. Plus - it's a system that people 'game.' Here are some links that cover the subject in more detail.

Alexa Totally Unreliable? — Business Logs
http://businesslogs.com/reviews/alexa_totally_unreliable.php

Why Alexa Ranking Alone is Worthless : SEO Book.com
http://www.seobook.com/archives/000034.shtml

Faking Alexa traffic ranking
http://www.techdigger.com/faking-alexa-traffic-ranking/

What are people seeing in the charts - is it real data, is it the site owner and all his friends boosting their site - is it a room of computers set to refresh and browse a site all day every day with the toolbar on? How would you know?

Should you care? YES!

I think it's getting to be time to expose how bad this is by having a bunch of people pick a random site each month - and have those people use the techniques for abusing the system to send that sit up the ranking. If the system is as poor as it seems, we should be able to manipulate the results and place the chose site near the top - at least in the under 100,000 range. Even better - we could put up a bogus site with a domain - and run it as the testing ground since it would start fresh with no data in the Alexa system.

Why challenge the Alexa Ranking?

Because we desperately need data like this - we need charts - we want to compare... and we use Alexa because it seems like the best thing out there. Which is why people are building and launching sites like http://www.alexaholic.com/ - a site I was just reading about over at Ajaxian: http://ajaxian.com/archives/alexaholic-get-your-alexia-stats-fix/

I can understand why the developer built the site, but I'm also troubled that someone who is obviously smart and intelligent would get behind such a broken system without posting anything on his site about it. Although... that would cut down on the uptake of his service - or maybe he doesn't know how there's a whole scene built around manipulating Alexa. Chances are, it's just the developer meeting a need and figuring it's the best thing out there and people want it. Which is ok, that's a developer's job.

I just have a hard time with cheating, and when you realize that cheating is going on in a system, and others rely on that system and do not realize the level of manipulation, it's a sad situation. False perceptions based on false data which in turn ripple out into other forms of falsehood is a receipe for disaster.

I'm not sure if it's possible, but I was thinking if Google Analytics could provide a feed of your traffic, that a 3rd party site could be setup that would allow people to register their traffic feeds and have that data analyzed against the whole collection. There's still a chance of people gaming their systems - but that could be dealt with to some degree on the analyzing side. Just check IPs and look for abuse patterns - which could be done through code and flag sites that seem to be generating fake traffic.

Or Google could just get some code together and simply add an option in Google Analytics you could check off that would merge your data into a pool for anaylzing and generating a Google Analytics account for 'the Web' - which would allow us to view data across all participants (limited view ofcourse).

Until the day comes that someone can provide a system that fixes the holes Alexa provides, we have to accept that there isn't a means for comparing a Website against another unless both sites provide their data untouched for analyzing to an unbiased 3rd party.

date: 03/20/2006 04:17:07 PM
name: Ron Hornbaker
url: www.alexaholic.com

I can understand why the developer ... someone who is obviously smart and intelligent

I'll take that a little out of context and simply say thank you. :)

The Alexa user base is only a sample of the Internet population

True, but it only takes a tiny sample of a population to attain statistical significance. Ask Gallup.

We all know it's IE users who have the Alexa toolbar installed, so that's a specific population - one that doesn't include all the Firefox users, or even the users of IE without the toolbar.

That used to be true, but now there are a significant number of Firefox users with the excellent Search Status extension installed, and that extension phones home your browsing data to Alexa, too.

or maybe he doesn't know how there's a whole scene built around manipulating Alexa

No, I'm aware of Alexa's shortcomings. I address them a little bit on Alexaholic's About page - perhaps you saw it. Is it possible to game Alexa's system? Yes for low-traffic sites, Maybe for medium-traffic sites, and No for high-traffic sites. If I'm comparing two top-20,000 traffic rank sites, I've got a pretty good level of confidence that I'm looking at data that is good enough for a rough comparative appraisal. And don't forget that Alexa is getting better all the time at preventing spoofing - especially these days, since they're getting so much more attention from the 2.0 and VC crowd.

Until the day comes that someone can provide a system that fixes the holes Alexa provides, we have to accept that there isn't a means for comparing a Website against another unless both sites provide their data untouched for analyzing to an unbiased 3rd party.

Well, that'll never happen. What I'd like to see is Alexa-like data, only from Google's toolbar (bigger N). Until that happens, Alexa is all us poor-folk who can't afford Nielsen data have, so don't knock it completely. Is it perfect? No. Does that mean it's useless? In my admittedly biased opinion, definitely not.

Gideon MarkenWeb Technologist & Electronic Artist

Gideon Marken
Web Technologist & Electronic Artist