I attended a conference of the Online News Association in San Francisco last month and heard USC Annenberg lecturer Dr. Dana Chinn present a good session on “finding meaning in the metrics” of Web analytics. (For those of you who are experts in Google Analytics, like my friend Jeremy Powers, this post is not for you. However, I will welcome all corrections. This post is for us beginners.)
Chinn began her presentation explaining some of the basic terms — and misunderstandings — used in Web metrics. For example, she started with the simple proposition: “Unique Visitors > Visit Websites > Generate Page Views.” Seems fairly straightforward, doesn’t it? But, as Chinn, explained “unique visitors” are actually “unique computers.” And therein lies a problem.
If one person accesses your company’s Website at home, then at work, and later from a hotel computer, that will count as three unique visitors. But, if one computer is used by four people (say, in a school) that will count only as one unique visitor.
Because of this, “unique visitors” can easily be over- or under-counted. Thus, Chinn pays little heed to these reports in and of themselves. As she put it: “So what?”
Ironically, publishers who are struggling to find homes online like visitor stats because they are comfortable to understand; they look like circulation figures from the ABC. But some online advertisers have become hip to the reality of these measures as have those who are calculating sales valuations of newspapers.
Now, on to “visits.” These aren’t as easy to find meaning from as they’d appear, either.
A visit is a period of activity separated by at least 30 minutes of inactivity. So, if a “visitor” clicks into your site at 1 p.m., surfs there for 20 minutes and then leaves the site, it counts as one visit. If that same “visitor” clicks into your site at 1 p.m., stays for an hour, leaves the computer for 29 minutes and then comes back to your site for one hour, that still will only counts as one “visit.”
But, if that visitor comes to your site at 1 p.m., visits for an hour, leaves for 31 minutes and then returns to your site, it will count as two “visits.”
With these caveats in mind, Chinn’s best advice was when she suggested that instead of relying only on the “counting” techniques, we try to use these measures to determine the level of engagement our site’s metrics suggest between us and “the people formerly known as the audience,” as Dr. Jay Rosen of NYU has dubbed them.
For example, try calculating the visits per weekly unique visitors. That could suggest if visitors are coming to your site with the frequency you need to build loyal relationships or not. Is your content engaging enough, she asks, that someone would want to visit more than two or three times per week? Hard to believe, I say, but the answer may just well be: “Honestly, no.”
Or, calculate page views per visit by week. When visitors arrive at your site, are they engaging with its content? Or do they land and then “bounce,” i.e. leave? Can you think why?
Maybe they didn’t like what they saw. Or didn’t find what they thought they would. Or couldn’t find what your site had “promised.”
Using metrics this way, I think, is both the logical and proper extension to help you evaluate the relationships you’ve now developed through blogs and other social media sites.
You can find a summary of Dr. Chinn’s presentation on SlideShare.