DM News SPOKE with Eric Peterson, CEO of Web Analytics Demystified. An edited transcript follows.
DM News: What are some key trends in the world of Web analytics?
Eric Peterson: I am tracking three major trends in Web analytics in 2007 and 2008: The impact of free systems like Google Analytics on the larger market, the difficulty of measuring Web 2.0 using a Web analytics 1.0 model, and the adoption of a process-oriented approach towards Web analytics.
According to data mined from some 12,500 distinct URLs tracked by Web Analytics Demystified, Google Analytics is being tried by nearly 25 percent of sites around the world, including 52 sites in the Fortune 1,000. While Google Analytics is good, questions still remain about how appropriate the latest release is for enterprise-wide deployment and whether Google Analytics is a suitable replacement for popular licensed solutions from companies like Omniture, Visual Sciences and WebTrends. I recently published a report that found that the majority of companies deploying free solutions are dramatically less likely to effectively use Web analytics and thusly must work harder to benefit from their investment.
Regarding the difficulty of measuring Web 2.0 using currently available Web analytics tools, many site operators deploying Web 2.0 technologies (RSS, AJAX, etc.) are unable to measure non-page-view events like subscriptions, drags, drops, and zooms using anything other than a historical visitor-visit-view data model. Web Analytics 2.0 is measurement strategies that reflect that our content deployment and consumption model has dramatically changed. Nielsen’s decision to standardize on time-spent metrics instead of page-view metrics is evidence of that.
Finally, the fact that Web analytics is hard is becoming increasingly obvious. Many good companies are starting to à take a process-oriented approach towards Web analytics: detailing specific tasks and sub-tasks that need to be completed so that the enterprise can really take advantage of its investment in Web analytics. Many companies are only now realizing that measurement doesn’t just happen and just because you measure doesn’t mean you’ll actually use the data.
DMN: How do search and Web analytics work together?
Peterson: Web analytics is able to measure the efficacy of search both externally (from Google, Yahoo, etc.) and internally (for sites running solutions like Endeca, FAST, etc.). The essence of the relationship is being able to associate visitor behavior and desired actions with the keywords and phrases that visitors use to find and navigate the site. Thanks to recent improvments in visitor segmentation technology, an increasing number of sites are looking more closely at searcher behavior and working to understand the relationship between branded and unbranded terms, paid versus organic search, and different relevance algorithms and presentation strategies for internal search.
DMN: Is the cookie still at the heart of Web analytics?
Peterson: Cookies continue to be an issue in Web analytics. The recent comScore report highlights that a relatively small group of “serial cookie deleters” is largely responsible for the dramatic inflation of cookie-based audience counts but also that many Internet users are actively deleting their cookies on a monthly basis. Despite this, I have still not seen a good alternative technology and so Web site operators still need to work to determine the impact that cookie deletion has on their reporting and adjust accordingly.
DMN: How can vendors better serve their clients’ needs in this market?
Peterson: I think the big gap in vendor-client relations exists in the services relationship. As an increasing number of organizations realize that Web analytics is hard, they start to address this by applying people to the problem. But experienced Web analytics professionals are both expensive and difficult to hire. The vendors have a great opportunity to help their clients fill the gap and use their services organizations to do far more than just implement code. In my experience, online forums and PDFs don’t solve the kinds of problems that clients have doing Web analytics; people do.