By Published on .

Web research is a "noisy" business. Claims and proclamations abound from research companies and the entities they measure.

However the same could be said about the growth of TV in the late 1940s and '50s, when scads of research suppliers offered many types of syndicated and special studies. Congressional involvement finally sorted through the noise, and only two of the dozens of companies survived and now only one has, ultimately, continued to succeed.

It's an axiom of research -- in any field -- that different methods yield different results. What is not self-evident is the definition of what is measured in digital media, how the universe is defined and which standards are implemented.


Three points come to mind: The appliance used to access digital media should be at the heart of the measurement; the definition of the measurable universe will affect what rating results are tabulated; and new standards are needed to accommodate digital media and traditional media measurement.

First, digital media as we know them are appliance-specific. The major appliance for access is, for the moment and for most users, a personal computer. There are a handful of users who have access from their TVs and from other small appliances, but such usage is limited.


Measurement that does not begin with the appliance is measurement divorcing itself from the medium, or rather like measuring TV without recognizing what a TV set is. Online services spawned the rapid growth of digital media over the last few years and these same online services are now the largest gateways onto the Internet. Ignoring measures of America Online, CompuServe Corp., Microsoft Corp.'s Microsoft Network or Prodigy is ignoring the largest accumulated audience of digital users.

Second, in digital media measurement the only way to come close to achieving the ideal universe -- the largest, most diverse and most inclusive -- is through measuring all digital media users (not just all Web users). It is accepted that two-thirds of all frequent PC users with Internet access do not access the Internet for weeks at a time. Advertisers and marketers may be able to reach this large target audience in many other ways: through proprietary online services, online/offline applications such as PointCast, free e-mail provider Juno Online Services and ad-supported CD-ROMs; and by driving this audience to the Internet more frequently in other ways.


However, only by understanding all digital media-user behavior, can it be influenced. To ignore part of it is a bit like projecting to a TV universe that ignores a major network and several cable channels.

Third, we need standards; but whose standards? Digital media are not like broadcast and print. Digital media are non-linear, and many of our traditional media measures cannot apply (once a TV show has aired, it's gone; Web sites and content within digital media change, but they are there all the time). Part of what makes the debate over audience measurement so intense perhaps is that it is easier to measure than any media measurement before it. Standards are still evolving.

We've gone from hits (one entry in a server log) just a couple of years ago to visits (a series of consecutive requests at one site) and page views (the number of pages requested and seen by all visitors at a given site).

As measurement has been refined, the industry has incorporated unduplicated audiences expressed in reach and projected thousands.


Frequency, the other half of the standard advertising equation, is a rougher measure (What is frequency? Sessions? Days? Pages? Total impressions?). The industry will debate this for some time. As for industry groups that will look to set standards, there are several involved now, including the Advertising Research Foundation, the Four A's Foundation, Coalition for Advertising Supported Information & Entertainment and the Internet Advertising Bureau.

The job of a measurement company is to help lead these discussions, but it's also patience. The so-called standards used in broadcast took nearly 20 years to evolve, and while digital time is faster, we are all learning of new ways to view this nascent medium.

Yes, it is noisy. Yes, there are challenges. But more standardized measures exist now than did two years ago, and there is clear evolution. What we do know, despite all the noise, is that for advertising and marketing models to exist and prosper there must be independent companies with syndicated measures. Consider other media: The better and more precise the measure, the more ad and marketing dollars have gone into the medium.

Mr. McFarland is senior VP-general manager, Media Metrix.

Most Popular
In this article: