MCN Guest Blog

Getting More Bang for Big-Data Bucks

How to Use Analytics to Understand Your Audience 1/11/2016 10:15 AM

In TV land these days, it’s a case of “data, data everywhere, but not an ROI in sight.”

 

We already know that TV viewing data can be purchased from leading data providers in the industry, and now some big guns like Comcast and TiVo have promised to open up their treasure troves as well. With millions of homes of daily TV viewing now at our fingertips, gone are the days when one could say TV is data-poor and CPM-rich.

 

In an industry where there is pent up demand for more transparency in ad performance, though, TV data comes at a huge premium. While conventional wisdom may suggest more is always better, a hefty premium imposes financial constraints on how much data one can buy. This is a real hurdle as the industry embraces data-driven and programmatic ways of ad buying and selling, which truly depend on this kind of data.

 

Which ratings currency are you using at the Bank of TV Audience Measurement? Some folks argue the current sampling isn’t really representative and are calling for a full census-based approach, which in my opinion wouldn’t yield much additional information.

 

The deluge of data into the TV ecosystem calls for a strategy that increasingly resembles Internet-style audience measurement, but the technology must be purpose-built for long-form content viewing. At the heart of the matter is what I would like to call data efficiency.

 

Let us now imagine an alternative scenario, in which we seek to learn about our audiences by putting advanced data science to work. Anyone familiar with the world of Big Data and analytics would agree that analyzing many dimensions of a viewer’s media behavior is always going to be more valuable than just counting how many people were watching a show. There is so much to be learned if dozens of disparate data sources — content metadata, geographic census data, consumer data, political data, social media data, cross-media subscription data and more — are combined and correlated through machine learning and predictive analytics.

 

With this approach, it turns out that with a fraction of the number of households of TV data, one can build an extremely accurate picture of viewing down to the individual and daypart. The resulting data efficiency, and the audience intelligence derived from it, are competitive advantages for those with the right technology.

 

The data science behind it is hard work, and it takes a dedicated, multiyear effort specifically focused on analyzing long-form content viewing. So, in the frenzy to license TV data, maybe a smarter way of going about it is to work with experts who have experience with the data and can make recommendations on the data mix and data science that fits your business and its audience intelligence needs.

 

Ask not what your data can do for you, ask what you can do with your data.

 

Pankaj Shroff is the founder and CEO of Psychability, a TV-analytics firm based in New York.

September
October

VR 20/20

The Times Center, New York, NY