Question on metering
I actually had meant to place a different post here this evening, but I received notice this afternoon that the Wicomico Neighborhood Congress meeting slated for tonight was postponed until next month. Something about snow?
Instead, I had a question I assume someone more conversant in the internet and tracking software could answer.
As most of you know, recently I finished up the 2009 Guide to the Maryland Blogosphere. Because there's just so many websites in our fair state, one criteria I used to cull the list was readership - I placed a lower limit I thought was pretty fair, although I did get some objections and that's understandable.
One thing I wasn't aware of until late in the process was the Alexa rankings. In theory, every site on the web with readership should rank someplace from 1 to somewhere around 29 million on their rankings. (My rank is just over 1 millionth. I guess that's not too bad.) I thought, hey, this is a way to judge readership where I don't have to ruffle any feathers!
But I saw some interesting differences between sites I found previously with fairly low readership that had relatively good Alexa rankings and vice versa. As a test, over the weekend I compiled the open Site Meter rankings or Quantcast estimates of a number of sites (56 to be exact) on BlogNetNews and in the Maryland Blogger Alliance.
For example, of the sample my site has the 13th highest SiteMeter reading (of 42 sampled), the 10th highest Quantcast (of 21 sampled), yet the 11th best Alexa rank of the 56. Yet another site with significantly better SiteMeter than mine (4th out of the 42 sampled) was only 13th on Alexa, and the site immediately ahead of me on SiteMeter was 38th of 56 on Alexa. WTF?
On the other hand, the site which was 14th of the 21 Quantcast sites was 5th on Alexa. Similarly, the 6th best Quantcast site was 31st on Alexa.
Now, it's my understanding that Alexa rank is based on prior 3 months, Quantcast is generally a monthly estimate (here my data is Quantified so it's supposedly more exact), and Site Meter is one week's worth of data. Obviously one could have a great week (like my Rushalanche) and appear to be a more well-read site than it truly is.
Another intriguing point comes up on the redesigned BlogNetNews site. They now go inside the influence numbers a little bit, posting a number of metrics for some top sites. (Go here and select Maryland. It's funny that someone forgot to change the gibberish fill text under the "About BlogNetNews Influence".) Selected blogs have their Alexa, Compete, and Quantcast rankings shown along with their authority measures from a number of services, including Technorati and Google. (I wonder if that's why I usually rank so well, because I get a number of external links - my average is third best.)
While I made this an extremely long-form question, it does pique my curiosity whether there's a good reliable method of determining readership. I'd like to have an easy method of determining if a site is well-read enough to belong to the Guide because it makes my life a little easier then!
Since I have the forum, let me also pimp my other interest. If you'd like to join the Maryland Blogger Alliance, let me know. We have well over 50 members - the listing on my site actually omits the latest joiners (it's a Blogrolling issue; hopefully this will be corrected in the very near future.)
Hopefully you'll pardon my digression from weighty political matters, but I had the opportunity to ask so I decided to do so.