Ratings

There is always heated debate over the quality and accuracy of the television audience measurement (TAM) system. After all, it is a statistical model based on a tiny sample of the audience cross-section whose ‘viewing behaviour’ is then extrapolated over the entire television viewing universe – to give broadcasters an assumed viewership performance of their channels and programmes on a regional and national level.

With any measurement system, there is room for error. In a statistical model, the scope for error increases – primarily because of the size of the sample and the behaviour of the sample being measured. The way tv viewership is presently measured is by installing people-meter boxes in a cross-section sample (of about 10,000) of viewer homes spread across the socio-economic food chain in pretty much the same weightage as the population data for these groups. Within an individual measured home, viewers are expected to record their ‘viewing behaviour’ on the special remote provided to them – behaviour such as who watched what, for how long, who left, who joined later, etc. Translated into English this means that the viewer being measured must actively record his or her own actions down to the last minute detail. And herein lays the contradiction. The best measurement systems keep observer and observed completely separate so as to not influence or corrupt the observed data. If the observed is also the observer, the resulting conclusions will always be, to say the least, flawed.

We now have an incredible opportunity to dramatically improve television viewership data – using the current push towards digitization. With the installation of set-top boxes, STB’s, in every home becoming mandatory, every home becomes addressable – which means broadcasters and government can keep track of and expect to receive their fair share of revenue from each home.

Going one step further, if every STB has an inbuilt software that is triggered by viewing actions as recorded by the remote control, linked to the EPG, and remotely collected – then without a viewer becoming aware, every single channel and programme change decision of theirs will be recorded. Since this software will be installed in every single STB, the sample size of measurement dramatically increases from the few thousands at present to the 30 and 50 and 100 million as STB penetration grows. This is a supremely robust sample providing almost real-time data with little scope of error or tampering.

Research purists would argue that while you would capture accurate viewership data of the home per se, you still wouldn’t know who actually watched what. But that is easily solved by marrying home data (available from the DTH Co’s) – number of people, ages, professions, location, etc – with the viewership data. In the beginning, the viewership information will be in absolutes, but with the additional layers of information, the data will become superlative. And truly real-time!

What’s better? A 10,000 size sample, extrapolated to tell you what, how and how many of 600 million people watch television…Or real-time data from 150 million homes that tells you specifically what, how and how many of 600 million people watched television.

The catch – making the STB smart enough to do this! With six DTH operators and numerous MSO’s, each installing their own type of STB, for an idea like this to succeed, it needs for everyone to act in concert – to gradually replace old STB’s, install the same common software in all the new STB’s, and then to outsource the data collection and analysis to a neutral 3rd party. The broadcaster and advertiser groups would be well advised to help make this happen, if only for their own good. It would benefit everyone – broadcasters, content creators, advertisers and platform operators.

It’s not impossible to do. It’s not terribly expensive either. The key – act in concert!