With Set-Tops, Who Needs Meters?

Over the last two decades, TV researchers have known that MSOs and advertisers have been missing revenues and losing on their return on investment. This is a direct result of the lack of accurate audience measurement for cable-television households.

Recent adjustments by Nielsen Media Research to address issues related to the lack of accurate cable numbers have been ignored, both by third-party data vendors and their agency clients. However, the real issue that's being sidestepped is the existence of a platform to measure cable viewership at a level of detail that's never been seen before. The power to do so lies entirely in the nature of the digital set-top box.

On any given day it is possible to know, through a secure and anonymous system, exactly how many digital households are watching a TV program at any given time. Current digital set-top boxes are capable of sending back viewing data that allows MSOs or independent third parties to calculate accurate ratings at the local or regional level — and eventually on a national basis as well.

Couple this with artificial intelligence, neural networks and certain inferential database computing expertise, and the delivery of demographics from those households — with the same or greater accuracy than the current Nielsen demos — is well within reach.

Consider the current issues addressed by Nielsen's total viewing source DVD, and one can clearly see that these issues can be solved completely — rather than incrementally —by moving to a system in which data is gathered from each set-top box and analyzed through some of the most recently-developed inferential engines.

Instead of improving ratings by lowering the weekly cume limit to 2.5%, the limit could be eliminated altogether, and the ratings for cable networks would be reported. Instead of expressing cable ratings with or without an alternate delivery system, they could be expressed on a much more granular and accurate region-by-region basis — or according to any other requested geographic combination.

In addition to getting accurate ratings data on a prescribed or flexible regional level that can be compared side-by-side with broadcast ratings, this new methodology creates an opportunity for a whole new set of statistics for measuring the effectiveness of video-on-demand content, linear programming and advertising flights.

Instead of merely counting how many people are watching in an average minute, one can use second-order statistics to measure how well each program or network holds its audience. Second-order statistics can be used to predict traditional first order statistics (i.e. ratings) in the same way that options can predict the value of stocks.

By being able to predict future ratings, MSOs would be able to accurately deliver the audience promised to advertisers, better select their VOD and linear content and generally have “sight” into an area that has been so far “invisible.”

These are just a few of the statistics that have been developed that will allow an understanding of cable-television viewing that is far beyond anything currently available.

The level of accuracy provided by these new statistics will give advertisers a greater return for their dollar and will allow MSOs to finally get their fair share of television advertising revenues.