Audience Measurement Is Catching Up on TV Consumption Habits


Eurodata TV Worldwide's One TV Year in the World report gives an annual overview of TV consumption and audiovisual landscapes in more than 100 territories around the world. In 2016, the daily TV viewing time is still solid with 3 hours per individual. To give even more accurate ratings in the upcoming years, TV audience measurement institutes are looking at ways to deliver a fully unified 4-screen TV audience measurement.

TV consumption worldwide

Worldwide daily TV viewing time: waning but still solid

TV viewing time calculated over the total population of 92 countries worldwide reaches the very round number of 3 hours per day per individual, a figure averaged over nearly 3.9 billion people. This is a 3-minute decrease compared to 2015, a very similar dip to the one observed last year. This decrease is a tendency shared by most regions in the world, and is stronger in North America and Oceania. However, a noticeable exception is South America, where the average daily TV viewing time has significantly increased between 2015 and 2016. Although on a decline, 2016’s average daily TV viewing time is still very high and remains mainly steady from an evolutionary perspective of nearly 25 years. On a regional level, this long-term outlook shows that Europe is at a consistently high level and that South America is on the rise, while North America, the Middle East and Asia show more significant decreasing tendencies.


Time-shifted viewing on TV sets makes headway

In 2016, 39 countries measured time-shifted viewing in their TV audience measurement.

In the 30 countries where the split between live and time-shifted viewing is available, we note that time-shifted viewing brought in an audience hike of 7% on average in 2016, over the entire population and the entire day. This percentage has slightly increased over the last few years. However, it should be noted that, overall, most TV viewing remains live. Of course, the weight of time-shifted viewing strongly depends on the studied market; countries like the United Kingdom, Switzerland, the United States, Belgium, Ireland and

Canada show a far better performance in that respect. A similar analysis of young adults shows parallel trends. Although we note a stronger slump in adults compared to the whole population, this decrease mainly impacts live viewing and not so much time-shifted viewing, which remains steady.

The relative audience increase among young adults is therefore stronger than for the whole population; we’ve recorded an average audience boost of nearly 10%, which has been slightly rising for the past two years. This analysis shows that an important part of the TV consumption measurements are lost in the mix: TV content viewed on online screens.


Fully unified 4-screen ratings is a must

A new step in TV audience measurement is in the works

For nearly a decade now, TV audience measurement has been put to the test by new habits in TV viewing.

The development of TV catch-up services and hard drives that can easily connect to TV set-top boxes has led to the measurement of time-shifted viewing.

This is effective in an increasing number of countries:

39 countries measured time-shifted viewing in 2016, which is 8 more countries than in 2015, and several more has taken the leap in 2017 (Croatia, Serbia, Chile, and more).

However, while more and more online services allow viewers to watch television programmes on their devices (computers, laptops, mobiles or tablets), audience measurements have mainly been limited to those in front of their TV sets – until now. In response to their local TV market needs, many TV audience measurement institutes have worked on ways to pick up these additional figures, to be up to date in reflecting today’s new modes of TV consumption. This is what we call a fully unified 4-screen TV audience measurement. Without access to this key data, channels and platforms communicate about online viewing using the number of launched streams or the number of unique visitors, often adding data over different episodes of the same show.

As a result, the online figures are sometimes impressive, but are in no way comparable to the TV audience figures the market is actually dealing with. Let’s keep in mind that the most commonly used “number of viewers” figure in TV is weighted in proportion with its viewing duration, which is quite different from the number of launched streams.

So, the first aim of 4-screen TV audience measurement is to provide comparable figures between audiences on different screens. Then, of course, the goal is to provide the market with the same level of precision that is available today for TV-set audience measurement: a variety of indicators (ratings, market shares, reach, coverage, etc.) and targets (by age, sex, socio-economic status, etc.).

Tapping into comparable TV audience figures on all screens will allow the market to:

Bearing this in mind, the good news is that 4-screen TV audience measurement became accessible to the market in 2016 in some countries (such as the Netherlands and France) and several more have it in the cards for 2017 or 2018, starting with Denmark (launched on 1 January, 2017), the US, Sweden, Singapore and many more!

How does it work?

Let’s start with a recap of how TV audience on TV sets is measured. In most developed TV markets, an establishment survey (often updated on a yearly basis) provides us with detailed information on a specific population regarding their TV equipment, TV subscriptions and platform availability depending on their gender, age and revenues. From this data, a TV panel is drawn up to represent the global TV population.

Then each household in the panel is equipped with an audience meter that records TV viewing by each panellist. This data is then automatically transmitted to the TV audience measurement institute that projects the TV consumption over the entire studied population. This process allows TV audience figures to include live viewing as well as time-shifted viewing (with some technical adaptations) given a great variety of indicators, such as ratings, viewing times, and coverage, as well as a breakdown according to demographic group.

TV consumption across online screens is far more spread out than viewing on TV sets alone, which makes it difficult to get consistent and accurate 4-screen ratings using only the panel process.

That’s why the first step in the release of consistent 4-screen TV audience measurement is for online platforms to collect “site-centric” data, or “census data”. Unlike panel data, site-centric data gives extensive information about the number of launched streams or the viewing duration, but it comes up short when tracking the exact viewer profile (age, gender, revenue, and so on). With site-centric data we can divide the accumulated viewing duration of a video by its length to get an average number of complete viewings. This ratio is very similar to usual TV ratings. Site-centric data thus provides us with ratings (in number of viewers) for programmes viewed on a live stream or on a replay platform, with a breakdown by screen, but not by target group.

This is why the second step of the 4-screen rating development is to mix site centric data with online panel data, which is thought to be representative of the online video consumer population. This merging between panel data and exhaustive data is called hybridisation, allowing the 4-screen ratings to be broken down by target group, as well as target group percentage, just as usual TV ratings.

 First results of 4-screen TV audience measurement

France and the Netherlands are among the most advanced countries for online screen measurement of TV programmes. Major TV channels are already measured on all 4 screens in these countries and the first feedback on online screens (computers, tablets and mobiles) is out. In France, 1 out of 5 people watched TV on an online screen in December 2016, while in the Netherlands, this proportion reached 50% of the population over the course of the same year.

Here’s another interesting insight from these online screen results: the audience profile on online screens is very different from the one on TV sets. Although young adults represent only 17% of the TV-set profile, this number rises to 50% for the online screen audience, in place of seniors. This result was to be expected, but it’s now confirmed using comparable data!

Florent Carême, Head of Research,Eurodata TV Worldwide

Back to top

Sign up to our newsletter

Are you happy to continue to receive emails from RTL AdConnect?
If so please fill in your details below

* Required