Our Online Dialogue colleague and analytics & optimization expert Jules Stuifbergen has for Webanalists.com wrote an article on Web analysts as the big data scientists of the future.

Web analysts and Database Marketers all use segments, profiles, outcomes and customer data. Yet there seems to be a large gap to be in thinking and practice.

Does it ever get right between database marketers and Web analysts?

Rather than a generation gap, it seems to be a data gap to be. The data sets used are fundamentally different. Gary Angel, a respected Web analyst with a BI background articulate it sharply: digital data is fundamentally different because it is a data stream, not a table.

The difference explained

Data analysis done by classical database marketers and statisticians assumes that data is row-based. So: a row with a number of columns in a database or spreadsheet has a certain meaning. The more rows, the more information of the same kind.

A Web analyst is not used to extracting valuable insights from this kind of static data. Where is the behavior? Where are the events that lead to outcomes?

Digital data is often different. Digital data coming out of a Web analytics package are data streams. Each line (a pageview, a click event, a purchase) is not very relevant by itself, but it only becomes meaningful when you consider the entire clickstream. Moreover, this data is incomplete, and not 100% reliable.

Database marketers, in turn, are not used to analyzing this data, and are often uncomfortable with the unreliability factor of the data, even if it is small.

Originally posted on November 13, 2012 at Data Data Data.com