Alexa ranking - How does it work?

Updated: March 21, 2013

It's like magnets. No one knows really. The thing is, in the past few weeks, at least at the time this article was written, but not necessarily at the time it will have been read by you, my readers, which is now for you, like now, but a future for me, now, at the time I am writing this, back then, you get it? So yes, last autumn, I observed a spike in my Alexa rank and traffic reach values. Prior to that, I had almost three months of steadily declining values.

The thing is, my real, genuine website traffic has been climbing steadily for the past year, without being reflected in the Alexa rank. And yet, with the release of Windows 8, it halved, which is good, although the actual net income of fresh traffic is not that huge. However, the proportion of Windows users seems to have grown, thus affecting the rank. So let's see what gives.

The graphs and whatnot

Here's what Alexa has to say about my three months at the time of the Windows 8 release last autumn - some fancy graphs:



Taking the selected time frame, supposedly, my site now reaches 70% more traffic in the past week than it did just a month before, which is total bollocks. And that plateau between September and October actually saw a nice, gentle steady increase in traffic, mostly related to Linux stuff.

Here's what Google has to say - now, this is not an accurate metric either, since approx. 20-30% of my audience blocks Google services and/or Javascript, but the trend is very similar to what the site urchin reports:

Google Analytics

Server urchin

As you can see, the least dramatic change is obvious from the site metrics. The fact Google sees more, means more people less likely to block Google services are coming, hence Alexa is reporting a more Windowsy trend. The dips you see are normal weekend days behavior, when there's less traffic, since people tend to waste more time using their computers at the work place. Stands to logic really.

So this makes me wonder. What would happen if all traffic to Dedoimedo was coming from Windows users? How would that reflect in the rank? And what does the rank actually mean, since it heavily depends on the distribution of users and browsers and whatnot?

I recall a handful of older discussion on how Alexa used to be inaccurate when it was only using its toolbar to figure out the statistics, but they had since made various algorithm changes, which still seem to be not as precise as I would like them to be.

Going back to the Alexa graphs, in reality, it's just a few more people Googling my Windows 8 recovery and Metro removal tutorials, as they seem to offer the Windows users worldwide what they need and how to resolve their problems. But the leaps in figures are truly awesome. Yes, they are, indeed.



Do not get hang up on any lovely graph online anywhere. Numbers are numbers, and they can be manipulated easily, especially when you use partial facets of truth to present the picture of the reality. I could have claimed my site suddenly becoming more popular, which is only somewhat true. Yes, there's more traffic, and the steady increase has been there since pretty much ever. But the quantum spike is not real. Humble me. It's just a different proportion of Windows users coming hither to bask in my sagacity and read my superb tutorials.

I may not solve the magnets problem, but perhaps you will have learned a little more on how Alexa ranking works, and that just a fraction of extra traffic, composed mostly from Windows users can yield huge differences in your e-ego. In my case, an influx of perhaps 15-20% more traffic makes Alexa think I am suddenly appealing to some 70% more people, which means their figures stem from no more than 45-55% of my overall userbase. And guess what, 55% of my userbase is Windows. Ta-dam. So it would seem that Linux and Mac users, the remainder of my population, seem not to count. There, mystery resolved based on observation and speculation. P.S. If anyone really knows how Alexa ranking works, I'd be glad to hear.