min read

4 ways to transform traditional TV viewing measurement

by Robert Nicklas , 12.05.2016

Our lifestyles are changing. We are Connected Consumers and our fast-paced lives have created a need for convenience and personalization. When it comes to entertainment, connectivity has converted viewers into “media multi-taskers”. It has driven an evolution in the way we consume TV. Yet according to the statistics, TV reach is declining in all major European markets, particularly among younger viewers. But is traditional TV viewing really declining? Or is the industry facing increasing challenges in capturing and recording its reach? And, what does this mean for broadcasters?

Tackling the challenges of TV audience measurement (TAM) data

There are two main challenges to continuing to deliver a “gold standard” TV audience measurement via TAM panels. Firstly, as the number of TVs in homes grows, more sets are unregistered. Secondly, linear (or traditional) TV viewing has become increasingly individualized.

Here are the four strategies we have developed to improve measurement of traditional TV viewing to help broadcasters understand their audience and distribute more relevant content.

1: Improving the number of TV sets under measurement

More than one in ten TV sets are typically unmeasured in households that participate in our audience measurement programs. This is particularly the case for TVs in bedrooms and children’s rooms, and for multiple sets in single-person households. With the right methodological changes, we have increased net reach by up to 12%, particularly among the key younger target groups. Recommendations based on these numbers are more solid.

2: Closing the gap on “missing values”

In order to provide a more holistic TV viewing picture, households with measured TV sets are used as donors for those with unmeasured sets. In other words, their TV usage is transferred using the power of data science to households with unmeasured TV sets that have similar characteristics (number and location in the house of TV sets, number of inhabitants, their age, gender, education, occupation, etc.).

3: “Nearest neighbor” principle key to individual usage data

To ensure that this approach provides not only as complete a picture as possible, but an accurate one as well, the recipient and donor have to be similar in structure. They must share, as previously mentioned, common characteristics and be close to each other in makeup. Statistically speaking, they must be “twins” or “nearest neighbors”. With the donor effectively substituting for the household with the unmeasured set, the location of the TV that viewing takes place on is also very important. If the donor’s viewing takes place on a TV set in the children’s room this is allocated to the TV set in a child’s room in the unmeasured house.

4: Capturing the individualized viewing experience

The results we have gathered have varied dramatically by both program genres and target groups. This underlines the complexity of the procedure needed to understand and capture a highly individualized TV viewing experience. Complex as it might be, however, we must pursue the means for tackling incomplete data. To ignore missing values and accept an incomplete picture would be the worst case scenario. The pros of our outlined approach to closing the TV viewing information gap are transparency, no burden on TAM panelists, and the provision of respondent level data.


As researchers, we have a duty not just to “fill in the gaps” but to understand how, why and where they occur. We have to use and develop a systematic statistical approach to handle incomplete data in order to help broadcasters better understand their audience. The results are evident with “missing values” with regards to target groups and program genres being filled. This is not about boosting reach and net ratings. It’s a highly targeted statistical method for adjusting differential non-registration and gaps in coverage.

Please share your thoughts by emailing me at robert.nicklas@gfk.com.