Eric Peterson has written extensively on Web analytics, including the book "Web Analytics Demystified," and runs a blog of the same name at www.webanalytics demystified.com/weblog/. He recently spoke with BtoB about Web 2.0 and its impact on metrics.
BtoB: What's involved in Web 2.0 measurement from your perspective?
Peterson: Web 2.0 is a multiheaded beast. Some of Web 2.0 is as simple as Flash and Flex, technologies that have been around for years and that most site operators, developers and measurement solutions are quite familiar with. I'd put AJAX in that camp too. Other technologies—like RSS feeds, and weblogs and social networks—are a bit newer. We need to get a measure of the impact of these technologies to enable marketers to do business online.
BtoB: What should marketers measure when they measure Web 2.0?
Peterson: When the IT department puts together an AJAX application, it's a beautiful thing. But do you measure every visitor interaction, every slide, drag and pull? Or none of that? That's the central complexity—what to measure? And it extends to RSS [and other Web 2.0 technologies] as well. Marketers I talk to are asking exactly these types of questions. The answers are not always immediately forthcoming. You could simplify things and measure Web 2.0 largely in a Web 1.0 framework of visitors, visits and page views. It works. The problem is nobody looks at that type of measurement and feels particularly good about it.
BtoB: So are marketers even ready to move from Web 1.0 to Web 2.0 measurement?
Peterson: One mistake we make sometimes when talking about measurement 1.0 is that we feel like it's a pretty well- understood area. That's a fallacious assumption. There's still a lot of confusion. When you move to Web 2.0, the most important thing to do is thoughtfully consider measurements in advance of development. If you forget to consider measurement upfront, when you sit down and say, `OK, was this widget successful?' how will you know?
BtoB: How do you measure a concept like engagement?
Peterson: I think engagement is very complex but I believe we are closing in on a nonconversion metric that any site can take and apply to any Web 1.0 or Web 2.0 site. It's a measure of visitor engagement rather than a measure of page views per session. But you have to be careful. If my visitor engagement score is 30% higher, that doesn't tell me anything—I don't know if it's good or bad news. My visitors may be more engaged but very unhappy. All of which comes back to the fundamental problem [of any analytics]: the need for bright people to figure out what it all means.