Counting Viewers: The Search for Cross-Platform Measurement
The fundamentals of TV ratings measurement haven’t changed since market researcher Arthur Nielsen devised a mechanism for tracking radio broadcasts and extrapolating the sample for the entire U.S. in 1942. It is a field ripe for disruption, with media agencies and broadcasters keen to agree on a system that combines the demographics of panel-based metrics with the census-style volume of digital measurement.
There is a growing feeling that the samples for panel-based TV measurement are too small for accurate assessment of an advertiser’s return on investment when contrasted with the exacting big data culled from connected devices.
Given that Forrester Research predicts internet ad spending is due to overtake TV spending in the U.S. in 2016, the issue should be addressed sooner rather than later. Then again, web ad spending overtook TV in the U.K. in 2009, and there the TV and digital markets remain divided.
“TV trading across the world operates just as it did in the days of Mad Men,” says Jamie West, director of commercial development and AdSmart at BSkyB. He contends AdSmart, an addressable ad system, changes this. “We can compete on a level footing with the digital media that claims to be the only market that can offer addressability, targeting, reporting, and understanding return on investment.”
The Internet Advertising Bureau reported £4.6 billion of TV ad spending in 2013, compared to £6.3 billion in internet ad spending. West claims that with AdSmart, Sky is already competing for an £8.3 billion slice of the pie—a mix of TV and digital.
The trend continues: Britons’ appetite for watching video content on computers, tablets, and mobile phones helped drive digital advertising spend up 16.6 percent to a record £3.5 billion in the first half of 2014. Video advertising on the internet and mobile phones grew by 59 percent year-on-year to £202 million in the first half of 2014. Mobile video advertising alone grew 196 percent to £63.9 million and is now the fastest growing digital ad format.
Yet here’s the thing: online video views are on the rise, and advertising follows the eyeballs, but linear viewing is not falling equally dramatically.
In the U.K., the TV set accounted for 98.5 percent of total viewing in 2014, according to Thinkbox, with tablets and laptops making up the remaining 1.5 percent. The Broadcast Audience Research Board recorded 123 million requests to view TV programmes from BBC iPlayer on computer devices in July 2014, compared to the 5.2 billion requests to view on a TV screen.
As the marketing body and ratings agency representing U.K. broadcasters, respectively, Thinkbox and BARB may have vested interests, but pundits such as Enders Analysis concur. It predicts that linear TV will still be at 77 percent in 2020, topped up by another 8 percent of catch-up views to the main set.
“There has been considerable buzz about the rise of digital video advertising, a widespread perception that people no longer watch TV,” says Jill Hind, Enders’ COO. “We believe TV will remain the dominant way to view video for many years to come.”
Figures from WPP’s GroupM show that TV revenue is growing in real terms although the yearly increases are slowing. TV spot ads actually rose rise 5 percent in 2015. Online video ads were on track to rise 50 percent in 2015, but from a significantly smaller base.
“Online is clearly growing, albeit additive not cannibalistic to linear, and skewed toward younger audiences,” says Sanjeevan Bala, head of data planning and analytics at U.K. broadcaster Channel 4. “The issue is that digital allows media buyers to execute a brand campaign using more measurable means to target viewers.”
The industry challenge lies in developing a single trusted currency that can look at both broadcast/linear and digital/nonlinear pieces in a unified way.
“The entertainment industry has been slow to embrace big data and the analytic capabilities required to make sense of what we know certainly with such scale and usability factor to be able to truly harness it,” says John Hoctor (right), VP of analytics for Rovi. “There’s also been a huge gap in the market for a platform that can integrate and harness data from all sources—traditional TV viewing habits, OTT and VOD consumption patterns, and data obtained from traditional STBs [set-top boxes].”
“There is a growing gap between broadcasters’ and rightsholders’ need to understand their audiences’ habits and media buyers’ request for more standardised tools that can help them decide on which channel and when to buy advertising spaces,” says Alex Terpstra (below left), CEO of Civolution.
“There are different measurements for different devices, and we need to connect the dots,” says Denise Turner, chief insight officer of Havas Media.
It’s a complicated process, but one that BARB has set itself to achieve in Project Dovetail. “The future of measurement relies on a hybrid measurement strategy that draws on big data,” CEO Justin Sampson (below right) said at IBC. “We have to gather behavioural evidence but also harness new data streams to enhance the outputs our customers are used to receiving.”
Even though social TV is yet to emerge as a commercial force, research from WPP-owned Kantar Media suggests Twitter activity does indeed amplify TV ratings.
Its Twitter TV Ratings tools unveiled in October includes metrics like total reach for programmes on Twitter, tweeters who generate the most buzz, and minute-by-minute Twitter activity overlaid with live ratings.
As part of Dovetail, Kantar has installed a software meter in 1,000 U.K. households to measure desktop and laptop video consumption. This is being extended to include viewing on tablets in 500 of those. Some critics will critique this small sample.
“Device-based data has its limitations,” Sampson says. “It doesn’t, for example, tell you who is watching.”
BARB also plans to incorporate return path data in Dovetail and is investigating audio fingerprinting to capture attention to short-form ads in particular.
In this regard, Kantar’s recent acquisition of Civolution’s audio watermarking technology could be significant. Nonetheless, it will be at least 18 months before Dovetail is proven and available.
“The reality is that nobody has ever delivered a data fusion of this scale and ambition anywhere in the world,” Sampson says.
In the absence of a universal gold standard, a number of players have emerged offering agencies and broadcasters a degree of assurance when buying audience profiles on VOD.
TVbeat, a London-based startup, received $2.4 million in funding in March 2014. It collects data from digital, cable, satellite, and OTT into an analytics platform and boasts algorithms that take account of anomalies such as the miscounting of views when a device is on but no one is watching.
“Shifting viewing habits calls for shifting TV analytics,” CEO Laurence Miall-d’Août (right) says. “You’d be surprised at how many broadcasters don’t do anything with the data they own.”
She wants operators such as Virgin Media and Freesat to share their data with TVbeat in return for benefits including insight on how to upsell genre programming or diagnose churn. TVbeat also plans to resell audience measurement data to broadcasters, media agencies, and joint industry groups, sharing the earnings with platform owners.
The company has aspirations to be BARB’s data supplier of choice. “[BARB] likes our solution. It’s still a very open process and we just want to make sure we get a chance.”
There’s competition though. Market researcher GfK, which took a majority stake in real-time data analysis firm Genius Digital earlier this year, offers a similar platform to integrate multiple datasets for a single view of a pay TV operation.
“If we go to market with a TV currency backed by a small U.K. technology provider [Genius] we are much less credible than one backed by GfK,” says Genius CEO Tom Weiss.
Other tools are being brought to bear on select aspects of audience measurement.
Divesting its audio watermarking technology enables Civolution to concentrate on NexGuard Forensic Watermarking and Teletrax TV analytics.
The latter, explains Terpstra, is at the core of TV-synced ads solutions and has already been integrated with Xaxis in its XaxisSync solution and the AppNexus’ real-time bidding platform. It tracks more than 7,000 brands across 17 product categories and more than 200 subcategories to offer real-time data on competitors’ advertising activities.
“Advertisers have adapted to the big data revolution, but are often limited by legacy systems and TV industry red tape,” Terpstra says. “By adopting real-time TV analytics, agencies and advertisers can integrate campaign and competitive data into their decision-making processes to get better results. In short, Teletrax TV-Synced Ads enables cross-device marketing on a global scale.”
London-based StreamHub combines digital viewing data from the client’s own players with social media data to create unique audience profiles.
“We can deliver insight into who is watching at a channel and series brand level, and also understand what other shows they have been talking about,” founder Aki Tsuchiya says. “We can help marketing teams discern what digital channels are most effective. Audiences for Facebook and Twitter should not be treated identically.”
Broadcasters are keen to explore data that may give them and their ad teams fresh insights as a weapon to fight competitors, he says.
“Broadcasters see digital as a big potential differentiator, and because there is no unified currency this is becoming a very open business proposition,” Tsuchiya says. “If you can offer better-rated segments to your sales teams for any particular content, they can sell it at much better prices.”
VigiGlobe also mines social media conversations to offer fresh insight, but the French company says its approach is different compared to other measures of sentiment: its algorithms have been written to distinguish between subject matters and which take language tone into account.
Canadian-British startup CrowdEmotion goes deeper into the emotional side of behavioural learning by recording facial movement using webcams (as Skype does) on the front of a TV or mobile device. It also records heartbeats and changes in skin tones, as well as audio responses.
What's better: "Millions" of viewers as defined by sketchy science and unreliable data, or thousands of viewers as defined by accurate streaming logs? Old media continues to bank on the "millions," and that will eventually lead to their demise.
The lack of consistent metrics is blamed as a key factor hindering greater ad investment in online video. Plus, is Google really the Death Star?
The site's new Social TV Statistics shows which programs are being tweeted about and how often.