The Alexa Issue - A Problem & A Solution
If you are not familiar with Alexa, it is a tool that tracks and ranks the traffic levels of websites across the internet. How does Alexa gather this data? They get it from their community of users who have installed their toolbar. This toolbar then tracks all the sites a user visits and this ultimately makes up the data you find on the Alexa site.
Initially, this would seem like a great idea. Users get the added benefit of having sites suggested to tfhem based on their browsing habits and all the web in general is able to see a ranking of all the sites on the internet. It’s a win/win situation, right? Unfortunately…no, it’s not.
There are a few inherent problems with the way Alexa goes about all of this.
The toolbar only works on Windows machines
“Who cares?”, you might ask. Despite Apple only having a little over 6% of the market share, that’s still literally millions and millions of users browsing habits in a significant demographic that are being ignored.
The toolbar only works with IE6+
With the massive surge of Firefox in the browser scene, an even smaller percentage of users are being tracked.
It requires a download
So not only does it require you to be running IE6 or IE7 on a Windows machine, but people have to be convinced that it’s a good idea for them to download some companies software that ultimately hacks in to their browser. In an age where viruses and internet security at are the forefront of everyone’s mind, the Alexa toolbar has got to be a hard sell.
So what is a prime example of Alexa’s flaws at work?
Recently John Chow made a post that mentioned, among other things, his traffic levels for the month of February. He said he had 266k+ pageviews and 127k+ unique visitors.
I’ve been on my anti-Alexa kick for a while and decided to run a quick comparison on Alexa to see what Alexa had to say about John Chow’s site versus a site of mine, The Apple Blog.
The Apple Blog gets over 500k pageviews and 250k visitors per month with our traffic being on a steady incline since we launched over 2 years ago.
With that info on the plate, check out Alexa’s comparison of my site and John’s site.
That’s very interesting to say the least. Not only does it show John’s site as having massively higher amounts of traffic, but it shows The Apple Blog as declining in traffic. Both of which are simply not true.
See, the problem in this situation is that a very large majority readers of The Apple Blog (well over 60%) are Mac users. On top of that, the non-Mac users are mostly using Firefox. And then, on top of that, the low percentage of users that are browsing the site with Internet Explorer are most likely to not want to bloat of their browser with another toolbar like Alexa.
All of these factors combined mean that The Apple Blog ends up with hardly anyone browsing with the Alexa toolbar which, in turn, gives the impression that The Apple Blog has relatively low amounts of traffic.
So what?
If you’ve read long this far you’re probably wondering why this is even a problem. “It’s just one of many tools for measuring traffic.” Sure, you could say that. And in many cases it really is a non-issue. The problem arises when companies start basing decisions around the stats.
Businesses like Text Link Ads and ReviewMe place a very large amount of weight on Alexa rankings. Obviously they need some way to rank sites, but the problem here is that the tool they are using is largely skewed towards Windows IE users when not every category of site has that demographic as the large majority. This way of doing business is hurting both the end user and the businesses themselves.
An Alternative
So after all of this ranting it’d be fairly lame of me to not offer up a solution.
The problem with gather web stats over something like television viewership, is simply the massive size of the internet. There are just too many websites floating around for people to be polled about similar to the way Nielsen does television.
So how is this problem solved?
I believe there are a number of solutions. Fortunately, these solutions actually already exist…it’s just a matter of getting people to realize how awful and inaccurate Alexa stats are, and to move towards an alternative.
Some existing alternatives:
Compete.com - Compete.com is sort of the ultimate alternative here. They have the largest source of a ‘clickstream data,’ which basically means they sample the widest range of users. It’s a healthy alternative to Alexa in every way.
comScore - comScore is second in line in terms of accuracy. They use more than one source for clickstream data and also separate US and international traffic. The problem with comScore is its low sample group (only a little above 100,000) and it’s not easily accessible or free.
Hitwise - Hitwise is generally considered to be one of the top methods for businesses to gain traffic comparison data on competitors. Unfortunately it’s very costly and really just isn’t worth it for the average Joe.
The ultimate problem here with Alexa is that their sample group is extremely biased due to technology issues that Alexa have blatantly chosen not to fix. Compete.com really should be the new standard for traffic comparison. They’re very aware of Alexa’s problems providing accurate data and are actively moving towards a better alternative.
I hope (and challenge) all of the companies that are using Alexa data as a major business axis, to rethink their decisions and move toward something that’s more accurate and ultimately more beneficial to the end user. It’s simply foolish to continue making business moves around an obviously faulty tool.
