Traditional Ratings Companies and Audience Measurement
Clearly there is a challenge to accurately defining and measuring the audience of a particular item of media, distributed over the medium of the network, as it is, at best, inaccurate and misleading and, at worst, impossible. This applies almost universally, be it in a ticketed event in a closed room or on a TV platform or an online distribution network.
In practice, access to a closed room is fairly tightly controlled, and in many circumstances access to an online system is also fairly tightly controlled, and this does limit the error in estimating the audience somewhat. However, we run into problems when we look at the most familiar broadcast networks. We have no connection to the edges of broadcast radio networks, and to all intents and purposes, we have absolutely no way of knowing if an end user of a TV network is “out there” and connected.
For this reason a variety of estimation systems have been evolved, and they rely on statistical sampling systems where a representative group within the expected audience has an extra system put in place that connects back to a measurement platform. From the activities of this representative group, extrapolation can form a basis for the estimation of the “real” audience size.