Guest Article: N D Badrinath: Are Small Sample Sizes Killing the Media Star?

Questions - and a scalable, low marginal cost suggestion - from an ex-insider.

Are we asking the right questions about IRS, TAM and BARC? Audience eyeballs are the currency of media buying and selling. That's how advertisers can see the value of a media buy. As they say in retail, "joh dikhta hai, woh bikta hai". Television and print media account for nearly Rs 30,000 crore of ad spends annually. Why do they struggle to establish a good eyeball count?

TAM, for more than 15 years now, has been fobbing us off with a grossly inadequate sample size of less than 10,000 meters. China, with a marginally higher population and one unifying official language, has 48,000 meters. BARC promises to start with 20,000 meters, growing to maybe 40,000 or 50,000 somewhere in the future.

IRS is the largest readership study in the world. But it still struggles to give reliable results for a country as populous, diverse and fast-changing as India. Some publishers want to depend on ABC circulation data. But ABC only shows how many copies are sold - not who reads them.

The 800-pound gorilla in the room - inadequate sample sizes

Most of the issues with TV and print surveys are directly related to inadequate sample sizes. But increases in sample size call for crippling increases in costs. Therein lies the rub.

Meanwhile, that man in the corner - the consumer - is flitting from print to TV to internet to radio to out-of-home to in-store while making that all-important purchase decision. Media convergence has long been a reality. Comfort zones are fast becoming uncomfortable.

We must shake ourselves out of traditional research mindsets.

KPMG reports that India had 67 million internet-enabled feature phones and 44 million smartphones in 2012. Smartphone numbers will explode to 236 million by 2017. Sample sizes of several million are possible if we can use even a small fraction of these devices to capture data. Technology and software is already available to do so. Market researchers around the world are testing smartphones to capture data, including what people read. Why is India not looking at this?

Scaling up sample sizes at near-zero marginal cost

Here's one scenario. A network of consumers downloads a research app that engages, delights and rewards them. Rewards need not be monetary - the app could provide some useful services, taking care that they do not influence media behaviour. In exchange, network members share some profiling data. We could devise minimally invasive data collection modes through the app. TV, radio, cinema and in-store media can have digital signatures that the app can detect. There are cost-effective technology options for print media to generate readership data. Computing and data storage power are not a constraint. India has enough IT talent to design and implement the required eco-system.

Most crucial - security systems can be set up to conceal all personally identifiable data.

Potentially, we get sample sizes in millions. Reporting becomes far more frequent. Sample sizes can scale up at near-zero marginal cost. There will be a learning curve of a year or two, but the end result will be worth it. The challenges posed by immense size, diversity and dynamism of our country can be surmounted.

A closing thought. A 5 per cent drop in ad revenues for just two media - print and TV - linked to lack of data works out to Rs 1,500 crore. This is many multiples more than any state-of-the-art, very large sample size, high frequency reporting system calls for.

What are we waiting for?

(N D Badrinath has, in the past, led the research teams for both the IRS and the NRS. He also has extensive experience in media planning. He is one of the founding partners of Aqumena, a Mumbai-based marketing consultancy).