2 | THE STATE OF VIDEO OCT 20183 | THE STATE OF VIDEO OCT 2018 WHERE WE STARTED 3 IT’S COMPLICATED8 SKINNY BUNDLES AND NEW COLLABORATIONS 11 CAUTIONARY TALES 14 LIGHTENING THE LOAD17 AND THEN THERE WERE THREE21AMAZON 23FACEBOOK26YOUTUBE28 AND FINALLY… 30 CONTENTS GroupM 3 World Trade Center 175 Greenwich Street New York, NY 10007 USAAll rights reserved. This publicationis protected by copyright. No part of it may be reproduced, stored in a retrieval system, or transmitted in any form, or by any means, elec- tronic, mechanical, photocopying or otherwise, without written per- mission from the copyright owners. Every efort has been made to ensure the accuracy of the contents,but the publishers and copyrightowners cannot accept liability in respect of errors or omissions.Readers will appreciate that thedata is as up-to-date only to the extent that their availability, com- pilation and printed schedules willallow and are subject to change. WHERE WE STARTED 4 | THE STATE OF VIDEO OCT 20185 | THE STATE OF VIDEO OCT 2018 the end game of unduplicated viewing measurement across all screensand channels is at least a couple of trading years away. Even in the UK,BARB’s Project Dovetail, perhaps the world’s most ambitious efort, hasbeen delayed and will not be launched at the time of this writing. Andstill, Dovetail will not capture all video on all devices, only all televisionon all devices. What is BARB’s Project DovetailBARB is the joint industry body that measures UK broadcast audiencesfor TV programs and impressions for TV ads. It has a sample of 5,100homes that return daily data on viewing on the main TV screen. A risingproportion, currently reported as 60% of the sample, also returns PCand tablet viewing. BARB is exploring how to add panel members’smartphones. Separately, 30 online TV platforms carry code to sendcensus-level data to BARB, revealing aggregate viewing (streaming andon-demand), top channels, and top programs.“Dovetail Fusion” is expected to combine current panel and censusdata with data from smartphones when it launches this year. At frst thefused measurement will aggregate program audiences. Commercialaudience impressions are to follow in 2019. This will reveal separateand combined reach for linear and VOD.Participating video distributors are measured according to BARB’s“gold standard.” However, no “walled garden” platform presentlyparticipates. There is no barrier though, as any content distributed onSVOD can be tagged for BARB to pick up on its panel – if the platformenables it. This would include such platforms as Netfix or Amazon.(Netfix has no ad impressions to measure, and doesn’t need BARBprogram ratings.) BARB does not currently measure viewing on smart TVs or via dongles,which is growing in importance as viewers demand more contenton the “best available screen.” Unmeasured video viewing is called“unmatched.”BARB is also investigating the possibilities for measurement of return- path data from set-top boxes. It already reports total impressionvolumes for Sky AdSmart addressable TV inventory (but not separatecampaigns or spots).WHERE WE STARTED In the fall of 2017, GroupM published our frst State of Video. In it, wedescribed key trends:Declining in linear TV ratings, especially among younger audiences. Increasing data application to linear television. Continuing progress with addressable television. Growing OTT viewing.Increasing availability of skinny bundles from legacy and new distributors. The threat of “cord cutting” to TV economics.New competition in content development and acquisition,particularly in sports. YouTube, Facebook, Snapchat and Amazon challenging the video advertising status quo. The likelihood of one or more “landscape altering” megamergers. A year later we ask, “what happened” The answer is straightforward:more of the same but not much more of the same.There are no new signs of life in linear television. Ratings continue to fall.Even mainstays like the National Football League have lost popularity forreasons upon which no one can agree, but everyone has an opinion. And,no new ad formats have emerged to suggest a sudden transformationfor the better. The solace for those selling linear television lies withsustained advertiser demand even as audience supply has fallen;infating unit costs are the silver lining of audience scarcity.However, there is a diference between lack of new life and impendingdeath. The ranks of television advertisers are swelling with new entrants,mostly direct-to-consumer businesses that have exhausted all the reachand awareness “performance” media aford them. It’s reminiscent of thedot-com boom for television of the late 1990s. Maybe it will end betterthis time. One thing is for sure: Linear television is still perceived to beas efective as ever, despite the absence of granular measurement. The application of data to television shows up in two ways: frst, inhow the medium is planned by advertisers; and second, by the wayit’s bought. It’s abundantly clear use of data in the planning is mostprevalent, and that real market transformation will only occur whenadvanced audiences become the currency of the medium, subordinatinglegacy metrics to valuable but circumstantial evidence.This change seems inevitable, as “video everywhere/anywhere” wouldseem the only logical solution for advertisers. Inevitability and speed arenot the same thing, however. In even in the most developed markets,WHERE WE STARTED Linear televisionis still perceivedto be as efectiveas ever, despitethe absenceof granularmeasurement. 6 | THE STATE OF VIDEO OCT 20187 | THE STATE OF VIDEO OCT 2018 Worldwide, TV measurement is improving, but not fast enough. By2016, 39 countries measured time-shifted viewing, up from 31 in 2015.RTL’s ambitions for what it calls “unifed four-screen measurement” isqualitative (behavioral) as well as quantitative. Countries developing thisinclude Denmark, France, the Netherlands, the UK, the US, Sweden andSingapore. Early results from France and the Netherlands showed 20%viewing online in a month and 50% in a year, with under-35-year-oldscomprising 17% of the main screen audience and 50% of those online.The young may yet return to the main screen. They are already saturatedwith smartphones, and screen size is a binding constraint. Big screensencourage viewing, particularly for SVOD, YouTube and sports. In the USA, for the 2018 calendar year, eMarketer estimates thatapproximately 9% of the television advertising investment willcomprise some form of “advanced TV.” This may include addressable,programmatic and use of measurement currencies beyonddemographics (we assume), but we are suspicious of this number.Supply and use of addressable television in the US continues to expand,but it remains a very small portion of the market. Addressability frstrelied on the ability to deliver messaging to individual set-top boxes, butnow it extends to the addressable inventory on connected or smart TVs,which are the clear majority of new device purchases. This massivelyincreases the number of addressable households and individuals, but thebinding growth constraint remains the proportion of total impressionsthat may be addressed. In the USA the sellers of addressable inventoryare the distributors with two-minute local “avails” in each hour.There are other throttles and bottlenecks around the world. The growthof VOD and ad-funded OTT is making very little diference. As we havecommented before, there are few technical barriers to addressability,but many commercial ones.It’s possible, however, that AT&T ’s acquisition of Time Warner,together with the 2013 acquisition of NBCU by Comcast, may catalyzechange. These vertically integrated companies have a signifcant share ofhomes served and content viewed. Should they choose to make all theirowned inventory addressable, the future will have arrived. Comcast’sacquisition of Sky in Europe foreshortened this future. Last year, we defned add