Getting hit when you’re good

Paola Quintanar is the web production coordinator at BCBusiness Online.

As the web production coordinator, it’s my job to deliver the statistics—sometimes good, sometimes not so good—about our website.

Many of you know that a “hit” is one visit to a website, but the actual story of web statistics is more elaborate—and less boring—than you might think. Hits are only the starting point; it’s possible to dig deeper into minutiae. Browsing a website is very much like strolling through a casino: any way you turn, your activity is being recorded.

Five important questions I get asked every month are:
—How many people are visiting our site?
—How long are they staying on the site?
—How are they navigating the site?
—Are they clicking on the ads?
—What’s the difference between a visit and a visitor?

The root of the problem is this: How can we assign statistics to images, electricity, and pixels traveling around the Net? The gold standard is the gold standard not because gold is inherently valuable, but because people agree that it is. This logic extends to the virtual world, too: in the avalanche of data the web provides, you can’t measure everything of value. But everyone can agree (more or less) to analyze traffic in the same way.

In truth, it’s not so harmonious. In the past year, I’ve used five analytics platforms—Google Analytics, WebTrends, Omniture, Awstats, Webalizer—and, perplexingly, each comes up with a unique snapshot of what’s happening on our site.

The reason that they vary is that operating assumptions are necessary to distill order from the informational chaos, and each platform’s assumptions are calibrated differently.

They differ on questions like:
Is it a real person behind the computer? (It could be, but it could also be a robot, spider, or crawler.)

Is it single person using one computer? (Might be, but it might also be two people looking at the same computer, or an individual using two computers.)

Does one computer equal one IP address? (If so, forget iPhones, cell phones, and BlackBerries.)

Are the browsers accepting cookies and Javascript code? (This one is a mixed bag.)

It’s a confusing world of web statistics, to be sure. Remember to be nice to your web production coordinators: they’re the ones squinting into the hail of data, trying to make sense of it for you.