« QuixtarSucks dot com | Main | New Tools System »
November 15, 2003
Are Quixtar's "numbers" reliable?
By QBlog in
Those familiar with Quixtar are also familiar with the myriad of numbers quoted within the business. In fact, sometimes those numbers can be a bit overwhelming and people often simply accept them as accurate because it's just easier than trying to verify every single integer. Yet inquisitive minds should always be asking, "Is this number accurate? Is it misleading? Is it a complete fabrication? Where did this number come from?"
Today I discovered some numbers in a WIPO decision called "Quixtar Investments, Inc. v. Scott A. Smithberger" and find them to be somewhat misleading if not purely dishonest. In the fourth segment of the document titled "Factual Background" we find this quote:
Quixtar.com has been quite successful, as evidenced by its performance during the 1999 holiday season when it averaged 15-20 million hits a day and was one of the top five e-commerce sites in the world based on revenue.I do not know how they determine the top five e-commerce sites so I won't question that claim but I do question the 15-20 million hits a day claim and I'll explain why I question the validity of such a claim.
Context
First, let's put those numbers in the context of Web traffic in general. Taking a look at Nielsen NetRatings we find that only three sites on the Web recieved more than 20 million unique visitors all of last week. The top site, Microsoft, had more than 53 million visitors which averages close to around 7.5 million per day.
Definitions and Intentions
The confusing aspect of the Quixtar traffic numbers is the use of the word "hits." Many unfamiliar with Web traffic would read "hits" and assume that a hit represents an actual site visitor. Technically it makes no such representation. The problem is that the word "hit" is commonly used (however wrongly) by the general public in reference to actual site visitors. In a similar way people often believe that the words "Web" and "Internet" are synonymous when they are not.
Hits - A hit is technically a server request. A server request is caused by visiting a Web page. However, if that Web page contains images, auto-loading audio or some type of multi-media plug-in then a server request is made for each unique file on that page. For example, this page you are reading right now contains between 7-20 unique images (and maybe a few more) and a each image is a separate server request. So, by visiting this page you've generated around 8-21 hits. If you visit a page with 100 unique images (and many sites contain that many) then you've just created 100 (plus the html request) hits. Now can you see how meaningless "hits" are when measuring site traffic? And due to the irrelevance of "hits" many in the industry automatically assume that when someone says "hits" they really mean page views or unique visitors.
Page Views - A page view is most important to advertisers because it's usually how their ad delivery is measured. A page view is exactly what it sounds like, a viewing of an entire page. It does not account for each file or document pulled from the server, only the page as a whole. If you visit this site and then browse to three other pages on the blog then you've generated 4 page views.
Unique Visitors - A unique visitor is an actual person visiting the site and is only counted once within a session (usually one day). A unique visitor can view 100 pages and generate thousands of hits but still only be counted once. This is really the most important stat when judging the popularity of a Web site and for rating purposes it's the only one that matters.
Did Quixtar Lie?
Who knows whether Quixtar was trying to be misleading, lying or just clueless about their own site stats. It's quite possible that they had 15-20 million hits per day (though even that is unlikely) as defined above but I guarantee they did not have that many unique visitors and I'm almost positive that there weren't that many page views (not to mention the irrelevance of such stats for Quixtar in general). I'll let you answer the question about Quixtar's intentions on your own. I just hope that by understanding the context of some numbers it will allow you to be a bit more discerning about similar claims.
REQUIRED READING: Here's an excellent examination of Quixtar stats.
Comments
Suck the data down while you can. Sounds like scientology all over again.
Oops, sorry, meant to post that to the MLMSurvivor SLAPP suit posting. Sigh.
Now I know what each page has do darned many graphic elements. Takes forever to load.
Reminds me of a joke/game my friend and I used to play on our employer. He wasn't net saavy at all, but loved the web-site my friend made for him. And he loved to see the "hits keep on coming". So much so, he'd take the office out for lunch every time the site hit a new hit record. So, after work, my friend and I would go home, log on and keep hitting the refresh button over and over again, knowing we'd be getting free food the next day!
Unfortunately I didn't last long on this job as the boss was a complete moron (and rumored drug addict) and I just couldn't handle the office environment.
Posted by: dmm | November 18, 2003 1:27 PM