Why Alexa Sucks

Alexa.com and similar websites claim to be able to rank traffic to any given websites, but relies on a really stupid way to do it. For “general” websites like Facebook, Google, and the like I’m sure it’s fairly accurate, but for niche websites, it’s ridiculously inaccurate. The reason for this is that the services rely on toolbars or other add-ons that users must actually add to their browsers. Do you have toolbars on your browser?

Personally, I have the Alexa toolbar on my browser, but that’s me, and I have it as part of a broader toolbar for webrankings. I only do because I’m interested in my site and other similar sites. If I didn’t have my own website and look at this stuff, I wouldn’t have it. In fact, I don’t know anyone else who does. I honestly think the last time I even saw a third-party toolbar on a browser was when my parents unwittingly added one on, which I promptly removed and educated them on.

People who would add-on the Alexa toolbar are either other marketing people, or people who were tricked into it or didn’t know any better, and that’s why for general sites it’s probably fairly accurate. However, take TuneLab.com for instance. Our target audience and primary audience is 20-30 year olds who are extremely tech-savvy, who are the exact opposite demographic that would have a third-party toolbar, thus making our ratings artificially lower. And that’s only home computers, because most office computers won’t allow something like that to be installed, and Alexa can’t track mobile users either. Conversely, take a website like Hypebot.com, which is a web marketing blog. I’d be willing to say that 99% of their traffic has something like Alexa installed, thus inflating their Alexa ranking.

None of this really matters of course if you’re aware of how Alexa works, but sadly most people don’t. Alexa is careful to say it’s an “Alexa rank” even though it does everything to imply it is an accurate count. Personally I think they should do more as a disclaimer, but they don’t. I also think they should give website owners the option to “opt-out” to keep the more ignorant from relying on Alexa’s bad numbers.

In the end, if you must use Alexa or Quantcast or anything similar, take into account the demographic of the site you’re researching. Better yet, just ask the site if they’ll send you raw data. Most sites would probably fudge their numbers, but there are sites like TuneLab that will give you unedited screenshots directly from the trackers.

Tags: