Ilana Mercer, June 18, 2003

In short order, you’ll hear some compelling reasons to doubt the popular Alexa Internet ranking system. But first, don’t fret: WorldNetDaily’s cyberspace status is indisputable. The site is among the top 500 most visited Internet sites. It ranks 517th on the more popular—and seemingly more pliable—Alexa, and 434th on Ranking.com.


While few regular Internet users likely know or care about the Alexa Internet ranking system, clusters of enthusiastic users toot the toolbar and for very good reason. In ranking website popularity, Alexa purports to use site statistics (when it can get them), “snapshots of the Web” (whatever that means), and related links to rank order the popular Internet sites. But, as Alexa.com readily concedes, the Alexa toolbar is the most important source of their information. “The toolbar,” says the site, “is a program that users install into the browser,” and which gives information about “the usage paths of the collective Alexa community.”


The Alexa toolbar does much more than that. Once installed, it inflates website ranking, an effect that is most pronounced for smaller, less-frequented sites. Download it, and your Alexa website rank begins like magic to improve in leaps and bounds, even when your site’s traffic remains constant, which is what I discovered when, curious, I downloaded the toolbar. The rating for IlanaMercer.com was already unrealistically high, but from then on it only improved, even though the site’s statistics remained constant for at least three consecutive months.


Fore sure, IlanaMercer.com is a great site. Personally, I think it deserves to be rated among the top 70,000 sites in cyberspace. But as someone who is not prone to mythical thinking or self-aggrandizement, I know better. Put it this way: The message of peace, an unfettered market economy, limited government, and individual freedom is just not as popular in the welfare-beset and warring America as one would infer from the rank Alexa gives my site.


I had a good idea what was happening. My theory was confirmed when I posted Sean Mercer’s domain name to a venerable list, many of whose members are Alexa devotees. As a consequence of that one posting, Sean’s website rank improved an order of magnitude! Again, Sean is a superb guitarist and musician. I recommend his CD. But in the era of Shakira, Kid Rock, and a sea of incompetents, it’s silly to imagine that the sudden spike in the popularity of his website reflected a growing interest in complex, well-executed, progressive-rock compositions. The truth was that the well-tooled list members had helped Sean’s ratings rocket, even though his total monthly hits only went from 10,000 to 10,020.


The Alexa service doesn’t always know how many unique hits a site gets. What Alexa most probably does is to work out a likely distribution of toolbars in the general population of Internet users. Hypothesizing that one in every X Internet users downloads a toolbar, Alexa will then assign a relative rating to a website, based on toolbar data. Alexa thus has no idea (or so goes my theory) that the spike on SeanMercer.com correlated with a meager 20 additional hits to the site. What they see is the activity coming from 20 toolbars. The odds that 20 individuals, most with toolbars, strike at once are clearly not that high. Sure enough, after that statistically insignificant surge, Dr. Rock went for a while from three millionth to 460,000th or thereabouts, based on a traffic improvement of 20 hits!


I’ve not seen graphs that plot a site’s traffic alongside its alleged ranking, but, in addition to the site’s total traffic, which is indisputably the best gauge, a consistent performance on separate rating systems is probably a good way to approximate an authentic rank. In addition to its 30 to 40 million page views per month and close to five million unique visitors, WorldNetDaily ranks consistently on the gauges. Both Ranking.com and Alexa attest that NewsMax.com, Townhall, and Jewish World Review are trailing badly.


Naturally, the larger the sample of Internet users, the less likely distortions are to slant outcomes, and, hence, the more accurate the results. Indeed, Ranking.com and Alexa.com are in general agreement on the position of the huge news sites. Ranking.com has CNN at 15th; Alexa has it at 20th. MSNBC lags a little at 23rd and 29th respectively. Fox News is 171st on Ranking.com and 135th with Alexa, and the New York Times manages 121st to Alexa’s 51st. The “Newspaper of Record” is still very popular.


Excessive discrepancies between rating systems ought to raise suspicions about realistically unintuitive popularity claims, especially when clusters of obsessive users all download the tool and massage one another’s rankings.


You read this space to get a dose of reason, not rubbish. Alexa says IlanaMercer.com is the 68,000th most popular site on the World Wide Web not because it is, but because an abundance of people with toolbars are hitting the site, including myself, while working with the comprehensive links page. Alexa records the continuous toolbar bustle on the site, and voila.



June 18, 2003

CATEGORIES: Junk Science, Technology