Digital TV Europe
As consumers fragment across platforms, how can advertisers be sure of the value of their investment?
Today’s fragmented viewership is the
biggest issue plaguing advertisers. Gone are the days when TV networks could
rely on delivering a broadcast ad to the living room in primetime with the
whole family gathered around. Today, advertisers have to navigate myriad
devices, platforms and VOD in order to find their audience. However, many of those
platforms are giving advertisers better opportunities to target specific
audiences. Piecing all of it together is a significant challenge.
“Increased consumer awareness and
cross-screen fragmentation means that the development of data-driven TV
planning and buying is growing,” says Ed Wale, managing director, UK and Spain,
SpotX. “This creates an opportunity for publishers and broadcasters to capture
incremental digital ad revenue when making their premium inventory available to
brands and advertisers across screens.”
He says this is particularly the case when
partnered with granular data reporting that enables digital publishers to offer
one-to-one audience measurement. “The risk is that there are eyeballs that are
not being counted and credited towards audience sizes when it occurs on a
mobile device or outside of the same day, seven-day or even 30- day windows,”
adds Belsasar Lepe, co-founder and CTO, Ooyala.
Bridging the gap Multiple efforts are
underway to bridge the gap between the traditional sample-based methods for
measuring broadcast audiences and new methods for measuring multiscreen
viewing. As Tim Jacks, principal, strategy at Cartesian underlines,
sample-based methods, where data is gathered from representative panels of
viewers, tells us something that data from most multiscreen services won’t –
who is watching the programme. “Is it someone watching by themselves, or a
whole family? And it tells you what demographics those viewers come from.”
By contrast, online multiscreen data tends
to be about what is being viewed – a programme was watched for so long, at such
a time, and was paused three times, for example – but doesn’t tell us anything
about how many people are watching, or what demographic they fall into. For
SVOD services, that’s not usually a problem, but for traditional commercial
broadcasters it matters. There are technical and business challenges to
overcome. From a business point of view sign-off is required from a range of
parties - publishers, aggregators, advertisers and vendors of audience
measurement tools and the ad tech that sits in the middle.
“Each party has distinct business goals, which
in turn means a unique set of incentives when designing a universal measurement
system,” says Aditya Ganjam, chief product officer at Conviva. “In addition,
there is a broader shift from traditional direct buying, which has been a
mainstay of TV advertising for many years, to more efficient programmatically
enabled addressable advertising.”
When you layer this technology shift on
top of multiple stakeholders jockeying for position, it’s easy to see why
coming to an agreement on a single measurement system is a headache.
“The reality is that ad sales teams are to
all intents and purposes still printing money – legally. There is not yet an overwhelming
need to adopt standards that fundamentally change how business is done,” says
Lepe.
“An important difference from the current
siloed currency systems, is that the most likely agreement will be for a
cross-media planning currency,” says Bas de Vos, global director of audience
targeting and activation, Kantar. “Trading will remain as it currently
operates; in TV with TV ratings; in online on agreed CPM prices defined via
programmatic systems.”
There are initiatives to at least have an
agreement on the requirements for reporting specific metrics from the IAB and,
in the US, the MRC (Media Ratings Council). From a technical perspective, the
issue is scale. Panel-based measurement starts to break down when you need to
measure consumption of niche content on niche platforms.
“The industry needs a clear, single,
independent view,” says Wale. This is beyond the ability of a single
broadcaster, particularly when taking into account all the different devices TV
is now viewed on. ITV Hub, for example, is available on 27 platforms alone.
“Since TV audiences are typically much
larger than most online audiences, on a per programme basis, it follows that
hardening audience measurement solutions that can accurately and reliably
measure very large audiences that span multiple screens and the three-day,
seven-day, 30-day, x-day windows will take time,” says Lepe.
Kantar Media’s Focal Meter tracks online/
player activity on any device connected to the home network. Recently upgraded,
it is currently in operation in Norway and Finland, while work on deployments
is ongoing in Turkey and Hong Kong with tests being carried out in the UK and
Switzerland. A People Meter is also ready to go to market. The most fundamental
redesign for 20 years, according to de Vos, it will incorporate real-time data
collection to deliver real-time ratings for all TV viewed on the main screen –
whether live, time-shifted or VOD.
If a universally recognised audience
measurement covering multiscreen and time-shift is not reached, confidence in
the medium could be destroyed. “The main danger is the perceived lack of
transparency, which would result in missed opportunities for media publishers,”
says Jacks. “Last year we saw online brands take huge criticism from major
advertisers such as P&G, which stated that it has deep concerns about lack
of digital measurement standards and the fraud that comes along with it. If TV
wants to take advantage of dynamically-inserted advertising, and the
flexibility that comes with it, it needs to ensure that advertisers trust that
they are getting what they pay for.”
Ganjam also says there is a need for a
single trusted third-party measurement tool to be agreed. “Brands will have to
buy across multiple different platforms without a consistent mechanism for
measuring performance. This will make designing attribution models more
challenging. Furthermore, there will be a trust issue because it will be harder
for brands to trust the data coming from individual platforms, publishers and
aggregators.”
The transition from sampled data to big
data measurement of audiences is underway. Nielsen, comScore, and BARB all have
offerings, but not all of these are trusted. “We are in the ‘hardening’ phase
of the lifecycle of these solutions and it is not clear how much longer we’ll
remain in this phase,” says Lepe. “In that sense, we aren’t at a place where
there are good or best examples. There are a lot of works in progress.”
One of them is Nielsen which rolled out a
total audience measurement tool that combines panel and census data into a
single measurement tool. However, during the initial rollout, some publishers
pushed for a more limited release because they felt the data could be
incomplete, and might offer misguided direction, according to Conviva.
A single measurement tool still hasn’t
been agreed. Another work in progress is Project Dovetail, which is expected to
be delivered in September with the publication of multiple-screen viewing
figures across TV sets, tablets and PCs. The UK’s BARB has also started reporting
TV set audiences for non-linear programming and reporting of
dynamically-inserted advertising.
“Building on this, we are committed to
delivering comprehensive reporting of TV set usage,” explains Justin Sampson,
chief executive of BARB. “Our third strategic priority is addressing the growth
in unidentified viewing on TV sets – viewing more than 28 days after broadcast
and non-broadcast viewing. This includes viewing to SVOD services, such as
Netflix and Amazon. We can measure viewing to SVOD services in the same way
that we report our panel members’ viewing to non-linear programmes, but it
requires the cooperation of the service providers.”
BARB is also exploring how router meters
could be deployed in panel homes to deliver an aggregate level of viewing to
SVOD services without their cooperation. It is currently trialling two meters:
one from Kantar in BARB’s core panel and another from Nielsen in a
specially-commissioned panel. “Our ability to deliver comprehensive insight is
improved with the cooperation of media channels and the platforms they are
distributed through,” says Sampson. “This is the principal challenge, as we
have the techniques to measure new forms of viewing.”
BARB believe a joint industry approach is key
to providing equitable figures across all media companies. Its recently
consulted a range of advertisers, agencies, broadcasters and social media
platforms about their expectations for a cross-platform currency, from which
two points emerged. First, the joint industry currency (JIC) principles that
underpin BARB (independence, objectivity and transparency) should not be
weakened or compromised. Second, the addition of any channels/ platforms
shouldn’t impact on the integrity of data collection and reporting methodology.
In addition, the survey clarified three
areas where advertisers and agencies expect any cross platform currency to
delivery comparability: all reported viewing metrics should follow established
TV conventions for duration, viewability and verification; advertisers and
agencies expect advertising reach and frequency to be calculated for placements
that meet industry-agreed standards for brand safety; and advertisers and
agencies should be able to plan campaign reach and frequency into editorial
environments classified by genre/ programme and to verify campaign delivery.
Conviva calls for a census-based approach
and points out it has the largest independent census data collection and
measurement network in the world. This has been deployed, it says, on behalf of
many live and VOD publishers and broadcasters “to measure QoE directly from the
consumer in-screen experience as well as very detailed engagement data in a
continuous, census-based approach.” Ganjam adds: “Most legacy-based approaches
were discrete event-based web counting or statistical sampling-based audience
surveys. Conviva’s approach is real-time true viewer-level continuous
measurement.
It was purpose-built for video, and the
sensor technology computes consistent and accurate metrics over a diverse set
of streaming devices and application software frameworks.”
Packaging for SVOD Once you have a
consistent cross-platform measurement tool, you can use data analytics to
extract deep insights about customers’ preferences, helping publishers identify
the sub-genre of content to commission in order to appeal to valuable
customers. The information could also be used to package and promote a SVOD
service depending on the target market. Netflix and Amazon both view data as
encompassing more than just engagement data and collect data everywhere along
the content lifecycle and user journey.
This allows them a form of content ROI
that no other players can match. Crucially, they also benefit from a closed
ecosystem and significant investment in data analytics.
“With these insights, they not only know
which assets to recommend to a given user from a given catalogue, but also what
content to create and add to that catalogue in the first place,” says Lepe.
“This approach has fundamentally changed the nature of how you operate if you
want to compete in SVOD.”
Traditional pay TV operators have made
progress in developing the use of analytics, driven by their increasing base of
connected devices. Jacks at Cartesian says the company has worked with multiple
operators to apply analytics to STB and device data “to develop a much more
sophisticated understanding of their cost base – which content is delivering
good value for money, which content is key to customer retention and so on.”
Arguably the best method of quantifying
how audiences are consuming video is simply to get the right information from
users. “Analytics benefits from audience measurement, not the other way around
as the resulting data can then be run through algorithms to better market SVOD
services,” says Jacques-Edouard Guillemot, SVP, executive affairs, Nagra. “The
operator wants to know which shows are being watched, when and on what device.
This creates opportunities, for example, to cluster subscribers very precisely
to recommend certain packages of content. If a user likes watching golf, the
service provider can suggest and build packages that will appeal to this
demographic.”
This is core to Discovery’s attempt to
better segment Eurosport’s fanbases. “Data analytics is already helping us
deliver content and marketing in a more tailored and personal way to fans, with
the goal of providing them with the ultimate experience around the sports they
are passionate about,” says the firm’s head of international research, Vincent
David.
The reality is that there are multiple
‘standards’ for online audience metrics and they are not always agreed on, nor
controlled by the companies involved. De Vos believe this “always online” world
will grow towards a set of metrics that the industry agrees on, but it also ensures
proprietary data is only available for the campaign at hand and can be
configured to expire upon campaign completion.
Kantar says it is noticing increasing demand from clients from
expecting the analyst to not only answer, ‘what happened’ and ‘what was the
impact’ but towards being more predictive: ‘What’s likely to happen’. “That’s
the space that we also hope big data and advanced analytics will fill.”
Impact of GDPR
Much has been made of GDPR and how it will be the end of services
and monetization built around extensive, per-user data and behaviours (like ad
targeting). The new legislation means users have to give explicit consent for
the PII (personal identifiable information) to be processed. This is
particularly relevant within the digital sphere when working with PII as a
device ID, a cookie ID or an IP address.
“Strengthened privacy practices will reap long term benefits for
the whole advertising ecosystem,” believes Wale. “At the same time, privacy
concerns could affect the ability of all parties to efficiently transact at
scale.”
Kantar cites the lack of access to audience data in walled gardens
and the privacy complexity driven by e-privacy legislation “are undoubtedly a
good thing, but it makes our work more complex”. Part of its testing is around
blockchain technology.
SpotX developed ‘Audience Lock’ as a kind of ‘data escrow’ to
conduct data-driven transactions so that buyers and sellers can maintain
complete control of their data sets. According to Wale, this alleviates
concerns by matching user IDs and audience segments in a controlled
environment. It also ensures proprietary data is only available for the
campaign at hand and can be configured to expire upon campaign completion.
“Measures like these can be valuable to the broadcaster who has a
duty to keep its data safe before for legislative reasons, whilst also ensuring
it maximises revenue,” he says.
Ooyala predicts that over time these restrictions will encourage
more transparency around how data is used.
“It is not hard to imagine a world where our personal data more
explicitly becomes a currency that each of us uses to access ‘free’ services
online (and maybe even offline),” says Lepe. “And does that mean a potentially
smaller audience? Yes, potentially somewhat smaller as some - likely mostly
older audience audiences - will opt out. But, I’d wager this won’t be a
materially large number.”
Conviva similarly believes GDPR and similar legislation, is good
for the industry. “This additional layer of security will help engender trust
among consumers, which has been challenged in the wake of recent data-privacy
scandals,” says Ganjam. “In terms of impact on a
company’s ability to capture data, it doesn’t prevent the collection of
audience behaviour data, but does ensure that consumers are fully aware of what
data is being captured and how it is being used. Ultimately this will help weed
out bad actors and should prove beneficial to the industry.”
