Two days in Singapore: charting a future for video measurement

asi’s Research Director, Richard Marks, reviews the 2018 APAC Television & Video Conference

At the time of writing it seems unlikely that the US and North Korean leaders will indeed be meeting in Singapore in June, despite the summit being announced whilst asi was in Singapore this May. However, our own video measurement ‘summit’ did go ahead, featuring senior opinion leaders from broadcast, television, online video, planning and advertising.

The APAC asi Television & Video Conference is now in its fourth year. Hot topics at this year’s event included the future of the video currencies (with or without the major online-only players?) online video effectiveness, programmatic and the ‘normalisation’ of SVoD.

Sally Wu of BBC Global News in Singapore

Content trends and SVoD in particular were the focus of our first session, chaired by Sally Wu of the BBC. Eurodata Worldwide collate currency TV data from all around the world, which puts them in a unique position to track the current health of TV. The global daily average for TV viewing time in 2017 stood at 2:56, down just four minutes year on year. However, viewing across Asia is lower, at 2:25, driven primarily by India and China. 44 countries of the 95 measured are now including timeshift in their ratings, but four-screen measurement is still in its infancy: sports specialists Global MMK felt that this is inhibiting their ability to track sports performance across platforms around the world.

SVoD services remain a preoccupation. Whilst SVoD services continue to grow their reach – with Netflix in particular doing carriage deals with cable and satellite platforms – analysis by Ampere indicates that time spent viewing to each SVoD service may actually be contracting. On average SVoD homes take 1.9 such services, meaning that they are now in a fight for attention. This adds further weight to the suggestion that SVoD, rather than replacing traditional TV services, effectively sits beside them as additional channels. So in that respect Netflix is arguably the new HBO as opposed to being the new Sky. You can hear more on this topic in a podcast episode here.

We got a detailed look inside Netflix as Nielsen presented an update of their award-winning paper from our Nice conference in November. Clearly the biggest Netflix shows are now getting an audience on a par with the hit shows on US broadcast television. Whilst delivered in a non-linear way, a surprising amount of viewing is concentrated on the first few days after release – binge viewing is not a myth but old habits die hard: viewers are keen to consume content as close to release or broadcast as possible.

This was amplified by data from Starhub in Singapore. Their Return Path Data indicates that whilst solus viewing of linear TV on their service remains high, solus viewing of VOD content is non-existent.  It’s clear that SVoD has had a major impact on TV around the world but, as ever, it supplements broadcaster content as opposed to supplanting it. The impact on bundled satellite and cable platforms however may be more profound.

In the afternoon of day one we moved seamlessly from content to advertising, with programmatic a key theme. Our main impression is that the hype around programmatic is starting to die down, to be replaced (thankfully) by a more realistic appraisal of what this approach can – and can’t – do for advertisers. In that regard there is a direct parallel with SVoD – part of the future of TV, but not THE future of TV. A paper from Mindshare quoted a prediction from Emarketer that 20% of all media will be bought programmatically by 2020, which is now just two years away. Google demonstrated that one of the pitfalls of programmatic is variable understanding of what it is and – in some cases – people assuming they know when they don’t.

Data from Think.TV in Australia indicated that the impact of ads on TV benefits from ads in full screen compared to sharing screen time on laptop and mobile feeds. Facebook research did not necessarily contradict this, but argued compellingly that mobile video works in a very different way to TV in terms of attention spans and therefore design, execution and impact. Also in this session we had – to the best of our knowledge – a first asi paper using AI to interrogate a database, in this case Marketing Scientist Group looking at the thematic content of online advertising.

The implication of programmatic  – as Mindshare summarised – is that it moves from placement planning as a proxy for reaching consumers to audiences segment planning. That undoubtedly has a place, but numerous papers at our Nice conference last November demonstrate that context still plays a vital role – not just brand safety, but the positive benefit of context – the value that the content and platform bring to the advert.

Meanwhile Ebiquity made the important point that simply chasing efficiency is not the answer: to increase ROI just reduce ad spend. Ad planning is certainly about delivering ROI, but doing so at scale. Marketing effectiveness, at least according to recent research by Binet and Field, has fallen due to over-use of digital and under-investment in television.  In all, it was an engaging session, expertly chaired by David Webb of Turner.

As is the asi tradition, day two of the conference put audience measurement very much centre stage. This year we took the approach of attempting to separate the science of video measurement from the politics. So, in the morning we focused on the latest developments in measuring video, with the afternoon looking at the appliance of that science – how the data is being used and in particular the barriers to total video measurement incorporating both broadcasters and the global social video players.

In reviewing their latest developments in audience measurement technology, the emphasis of the major TAM suppliers remains very much on using data science to fuse elements together, those elements most likely to be PeopleMeter panels, router meters and server data, supplemented in many markets by online panels when the TV panel alone is not large enough to measure online behaviour.

At asi we have been tracking the development of the SGTAM service in Singapore and that is now starting to produce fascinating insights on multi-screen measurement, particularly on real – as opposed to claimed – smartphone usage, in terms of duration and activity type. Smartphones are used during 13.4% of TV viewing but, looked at from the smartphone perspective, 22.5% of all smartphone usage occurs when the TV is switched on. So which medium is distracting the other?

With agreement on the continued importance of PeopleMeter panels as the ‘core’ of measurement, BARC in India demonstrated how connected technology can enhance the running of the world’s largest TAM panel, whilst Auditel in Italy showed how a Set Meter Panel can prove cost-effective in expanding the scale of TV panels, an approach that may be of interest in the APAC region where panel sizes are under pressure to cover Pay TV operators more effectively.

Yan Liu outlines TVision Insights’ eye tracking system.

Whilst the primary focus is on exposure, Kantar Insights flagged up the importance of attention in a crowded media environment and appropriately enough we heard from a new entrant to the market: TVision Insights uses eye tracking to measure actual screen attention and already has a sizeable US panel. The company pivoted early on from an approach that used audiomatching in wristbands because the devices overheated and could explode – a good example of the need for companies to be willing to fail in order to succeed!

On Friday afternoon BARC’s Partho Dasgupta chaired a session in which we looked at the latest developments from the region, which included addressable TV delivered by TVB in Hong Kong via Nielsen’s marketing cloud, Facebook and TV combined reach in South Korea and an insight into how Video Research is preparing to measure the 2020 Olympics in Japan across all available screens.

We concluded with a session focused on the politics of total video measurement. GfK, Kantar and Nielsen were joined by OzTAM and Google on a panel chaired by Partho. Two themes in particular dominated – should media companies like Facebook and Google be incorporated into video measurement currencies and what is the likely future of video currencies?

On the first point there did seem to be consensus on the panel and in the room that Facebook and Google should be part of a shared currency. However the barrier to entry would appear to be what metrics should be used. ‘Legacy’ broadcasters insist that online video companies should fall into line and agree to report ‘traditional’ metrics used to evaluate broadcast TV. This is seen as comparing ‘like for like’. But how ‘alike’ are online video and broadcast? We have seen compelling evidence that they work in different ways in terms of duration, optimal design and impact. Complementary, but different.

Yoshihide Ikeda (left) and Etsu Taniguchi from Video Research in discussion during a panel session.

It seems to asi that the industry really faces three choices – require online video to use broadcast metrics, allow online video to share the measurement system but output different metrics relevant to each medium, or reach compromise on a new set of shared metrics. At asi we suspect that the first option is highly unlikely. We will be providing a forum at our events that allows both parties to come together on the principles of a shared system and evaluate just how practical a shared system and/or metrics would be.

Despite these differences, the real step forward is that we are now having conversations in a constructive manner. It does now seem that companies like Facebook and Google are recognised as valued contributors at asi events, as opposed to perhaps interlopers viewed with initial suspicion by some delegates when we first started to invite them. This widening of the asi community is a useful base to build upon to take measurement forward.

As well as three possible approaches to Total Video measurement, GfK’s recent whitepaper outlined three possible futures for the industry currencies: Super JICs regulating all data, walled garden anarchy or complete decentralisation via blockchain. You can hear more about the scenarios in this recent asicast. There does seem to be consensus that the shared currency model is the optimum one, but agreement does need to be reached on how online video can best be measured and reported, or the currency model will be at risk.

Finally we have to note the absence of our CEO Mike Sainsbury at the event. Rest assured he is about to binge watch all the conference video and will be back at the helm for our upcoming International Television & Video and Radio & Audio Conferences in Athens on 7th– 9th November. Stay tuned for more information about those conferences and if you have any potential papers for consideration or topics you would like us to cover, please do get in touch.