From TV sets to viewers
To understand the role that measurement plays in decoding the relationship between ‘man and machine’, it’s important to understand how measurement has evolved. Before Kantar introduced People Meter technology, passive measurement of TV viewing had largely been conducted using set meters. A meter attached to the TV set in the home monitored which channels were being tuned to and produced a household-based estimate of TV ratings. This could be transformed into an estimate of actual persons viewing by simply applying a factor derived from the number of people per households.
The advent of the People Meter – which means that individuals, at home, can register their presence in front of the TV set – revolutionised measurement as it revealed that the relationship between content and viewers was not constant, varying greatly by day part and the type of content. This was a breakthrough for both programme makers and advertisers and it is no surprise that People Meter technology endures at the core of measurement to this day.
The ‘big data’ opportunity
Huge opportunities are now presented by the availability of server and census data for measuring consumption of streaming video, presenting a major step forward in granularity of measurement. However, as with the TV set, devices are not people. Without an understanding of who is in front of different screens and different types of content, there is a danger that measurement could take a step back to those days of measuring sets rather than people.
From devices to people
Kantar Media is leading the way in the introduction of cross-platform measurement around the world, with the combination of people meter and router meter technology enabling us to measure viewing across all platforms and devices in the home. Findings from our cross-platform services are reinforcing the critical importance of persons-based measurement.
For smartphone usage, the relationship between device and people viewing in most cases is 1:1 and the majority of the viewing is attributed to the device owner. This offers a high degree of reassurance that addressable targeted campaigns will reach the intended recipient.
When it comes to tablets and laptops/PCs, there also tends to be a single user, but that user is not always constant: tablets and laptops can often be used by different household members at different times of day and whilst some more diligent users may log out and log in as a different user, many do not. So, these devices can be more challenging in terms of targeting.
TV set usage is where that 1:1 relationship is transcended. By their very nature, TV sets are conducive to shared viewing. Recent data from BARB in the UK reveals that, across the broadcast day, on average 1.37 people aged 4+ are in front of each TV screen. This falls slightly to 1.29 during the day when less people may be about, rising to 1.49 in the evenings (18:00 -23:59).
Bearing this in mind, Kantar’s recent report ‘The Future Viewing Experience’ highlights a number of vital implications for the industry going forward.
Understanding the way that viewers actually watch television.
Firstly, there can be a tendency for technology manufacturers to forget that video viewing on TV screens is a shared immersive experience. Some past innovations that looked set to revolutionise TV viewing failed because they forgot this simple fact. 3D televisions, curved TVs and interactive technology assumed a single-user relationship between screen and viewer. Much viewing of television is essentially a lean-forward experience, with interactivity best left to video games. As the report highlights, TV technology succeeds when it amplifies how people actually watch television, as opposed to simply exploiting what is technologically possible.
Smart TVs mean larger audiences for the streamers
Meanwhile, as the report also confirms, the impressive growth of Smart TV adoption means that an increasing proportion of the viewing of streaming services is happening via the television screen, migrating from devices that had low co-viewing to larger, more shared screens.
Kantar Media data shows that usage of different types of VOD differs significantly by type of screen. Early data from BARB in the UK confirms that the majority of long-form VOD streaming is now on large, connected TV screens that are likely to have higher co-viewing than on mobile devices, where services like YouTube and TikTok attract the majority of their viewing. This shows the importance of measuring and understanding all audiences, whether on Smart TV or smartphone
This is great news for SVOD and AVOD platforms as it means they are being consumed more and more by viewers watching together, as long-form viewing migrates from personal devices. Barb data shows that the co-viewing factor for SVOD and AVOD content on TV screens is 1.49.
This shows why streamers are increasingly signing up to persons-based measurement currencies as they need to understand who is actually watching: the registered subscriber, someone else or multiple people? Also, as they move to ad-supported tiers they want to monetise a viewership up to 50% higher than their own device-based data is able to detect.
From impressions to impact
The fact that co-viewing varies greatly by screen has the most profound implications for both the buy and the sell side of the advertising eco-system. Integrated persons-based measurement allows the industry to move beyond impressions (sets tuned, or streams served) towards a measure of actual impacts across devices, to turn devices into people.
This imperative drives the importance of representative panels, measuring people across all screens. Cross-media ‘calibration panels’, also operated by Kantar Media in a number of markets, offer an opportunity to strengthen the consumer signal by calibrating large scale data sets and unlock audiences by delivering true reach and frequency across platforms.
Without people-based measurement, there’s a real danger of combining variables. At Kantar’s latest World Audience Summit, Marta San Pedro of Dentsu highlighted this:
" An impression is a device. On mobile it is approximately the same data users, but on big screens you have average 1.4 users per device so you cannot just say 'this is an impression, that is an impression’ and add them up, you have to go from devices to users and then add the profile to people. All over the ecosystem people are mixing the variables. You cannot just say 'this is an impression for TV, it's the same' – no, it's not exactly the same. In audio-video on big screens you have more users than impressions”
Panel-based measurement shows how co-viewing varies across screens, and who those viewers are. This means that the total viewing audience can be understood and monetised. Planners can understand which screens and type of content are more likely to attract high levels of co-viewing and which have a more direct 1:1 relationship.
Unlocking the relationship between the viewer and the screen.
Ultimately, people-based measurement is highlighting the importance for us to understand how screens relate to people: is viewing on personal devices a shorter, but more intimate, more targetable experience? Is viewing on larger screens more immersive but a shared experience? What does this mean for ad creative as well as targeting?