MEASURING THE EFFECTS OF INTERNET ADVERTISING: PART II
by Larry Coleman
In Part 1 of a two part series, I noted that the Internet initially was characterized naively as the perfect advertising medium. The industry has matured and realistic efforts have emerged to solve fundamental media measurement problems on the Internet. However, rapid evolution of the technology and rapid growth of the Internet have challenged our ability to keep step with the evolving measurement problem. Therefore, the current state consists of three imperfect measurement solutions. In Part 1 I Introduced the first of three methodologies -- Site-centric Measurement. In today's article I will discuss two other methodologies that are moving us toward a comprehensive solution -- User-centric Measurement and Survey-centric measurement.
To compensate for the lack of user behavior measurement in site-centric methods, a class of User-centric Measurement methodologies have emerged. In User-centric measurement, a panel of users agree to install monitoring software on their PCs that reports every site they visit as well as how they use their computer off-line. In addition, these users provide baseline geo-demographic information. The output is a snapshot of Web usage in time. Site-to-site comparisons are possible because each site is measured the same way and the results can be projected to the universe represented by the sample.
Measurement companies like Media Metrix, NetRatings, RelevantKnowledge and Neilsen Media Research conduct active user-centric measurement panels. However, because the Internet is fragmented into hundreds of thousands of sites, one of the limitations of user-centric measurement is that the findings can be projected only to the universe represented by the sample.
Advertisers are critical of the panel selection methodology and the scope of the selection. Business users, who represents nearly 35% of Internet users, are either underrepresented or not measured at all. Consequently, there is some significant difference in generic web measurements, which cast doubts on respective proprietary measures. A further limitation is that measurement of changes to the feelings and attitudes of the user, a bulwark of traditional media measurement, is generally lacking. Such measurement is equally essential.
To deal with the effects of fragmentation in user-centric approaches as well as the lack of traditional media metrics, a class of Survey-centric Methods have emerged. Fundamentally, the methods have to do with having users complete surveys that ask them detailed questions about their feelings and attitudes, much as is done in traditional market and media research. The surveys can be simple surveys or, they can have more of the flavor of a research study such as those conducted by Harris On-line or MBinteractive.
Such surveys could be used to augment site-centric and user-centric measures. Furthermore, if the surveys are designed with traditional media measures in mind, it becomes possible to effectively compare different media. Understanding, measuring and valuing the Internet investment according to traditional media measurement criteria is necessary for determining Internet specific and overall media ROI.
The Internet advertising industry is making great progress toward a comprehensive solution. The ultimate Internet measurement solution must incorporate critical elements of all three methodologies. Fortunately, the technical underpinning needed to achieve the objective is inherent to the Internet.
---
Larry Coleman can be reached via http://www.virtualgold.com