Thanks so much for the write-up. With respect to the methodology, I can outline how we approached it. We surveyed 14,000 real website visitors across a broad portfolio of media sites. In essence, we wanted to look at content-based publishers who ran multiple types of ad units on their sites. We used out standard, two-stage survey methodology, and tracked self-reported likelihood to click on a variety of ad units. We employed a multiple-select type question, so that we could identify collinearity between ad units. Behavioral tracking on those sites would certainly have furnished some interesting results, but it would not have yielded visibility into visitors’ web-wide ad engagement patterns.

Thanks!

Michael Whitehouse
Senior Marketing Analyst
iPerceptions