Last month, I had the honor of speaking at the TV of Tomorrow Show in San Francisco, where I participated in a panel discussion about the future of TV.
What is TV?
I am not a big fan of using the word TV in the first place. Instead, it’s easier to talk about video, which takes all kinds of forms – social and user-generated video, short-form video, long-form video, live video, sports, video advertising, e-sports – and video content may be consumed on-demand, through broadcast, at various speeds, and across a number of screens, including smartphones, smartwatches, traditional or smart TV sets with or without auxiliary devices (such as gaming consoles), tablets, and an increasing range of other devices. Since the launch of YouTube thirteen years ago, video viewing has transformed more than it changed over the previous five decades: consider the launch and evolution of services like Snapchat, Facebook Live, ESPN’s new digital video streaming services, Hulu, Netflix, Amazon Prime, DirectTV, and Twitch.
Problems with Today’s Methodology and Measurement Standards
Onstage in San Francisco, one of my main goals was to educate conference guests and generate a thought-provoking discussion around the measurement of cross-platform video. From Verto’s point of view, one of the primary concerns in this space is methodology. How do you develop a way to compare digital and traditional linear/time-shifted TV, and which single-source, panel-based approaches are needed to assess the consumer across key media platforms and measure different platforms on a comparable scale? While we prefer and have executed on the panel-based approach, the challenge, of course, is one of scale; panels work well to generate consumer insights and ratings, but they’re not always capable of delivering advertising tracking based on typical sample sizes, for example. And while it’s tempting to simply build a bigger panel, if you focus on just one single-source panel rather than multiple smaller panels for different devices or platforms, you’ll eventually have to use fusion to bring them together. While fusion may be able to deliver high-level ratings data, it falls short when it comes to producing practical data for day-in-life modeling, journeys, path-to-purchase cases, or any deeper consumer insights.
Many companies also use census data. While it plays an important role in panel stratification and weighting, the practical reality is that big digital players, like Google, Netflix, or Facebook, do not allow tagging (you will be left with only a partial world being census-tagged), linear broadcast cannot be measured (despite the rise of digital video, this will still represent a significant part of TV consumption for the next decade) with census data, and data protection efforts such as GDPR limit the use of tagging, tracking pixels, SDKs, and other tracking technologies.
Taking On Today’s Standard: Nielsen
Another major conversation around TV measurement concerns its existing standard of currency – Nielsen. As we all know, Nielsen’s metrics are based on average audiences, giving advertisers a metric for the average number of people who view an ad during a particular unit of time. Oftentimes, these metrics are averages across particular demographics, and provide an easy way to compare audiences between shows. Nielsen is also accredited by the Media Ratings Council (MRC), and has a large organization dedicated to normalizing the data, correcting anomalies, and ensuring consistency over time. But despite these measures, the market still feels that Nielsen’s methodology has weaknesses.
Nielsen has fine-tuned their methodology over the past five decades, benefitting from a legacy of trust and the household-name status of the Nielsen brand. They also can demonstrate stability with their approach and numbers. But consumer behaviors are changing quickly when it comes down to TV viewing. Based on discussions I’ve had with many of Verto’s customers, complaints about Nielsen typically boil down to one of the following:
- Granularity: Nielsen’s approach is largely based on households and not individuals. They also still rely on an approach largely based on rating audience demographics, size, and frequency/engagement. Many of today’s clients expect solutions requiring greater depth; for example, an approach that understands attribution and what happens before and after an exposure.
- Scalability: The scalability of Nielsen’s panel, which is based on hardware based TV set box meters, presents an issue. Some clients have use cases requiring much larger sample sizes or samples capable of measuring programmatic advertising or more advanced advertising attribution tracking.
- Flexibility: Nielsen has struggled to measure digital devices, or OTT (over the top) content, on a comparable scale to traditional TV. Out-of-home measurement has also presented a challenge, although Nielsen is now leveraging PPM-type (Personal People Meters) solutions used for cross-platform 24/7 metering (via a meter from Arbitron).
Why We Need A New Standard to Measure TV and Digital Video
From Verto’s perspective, measuring digital is a much different game from traditional TV. First, there are a handful of giants like Youtube, Amazon, Facebook, Snap, Twitter, and Google, who already have access to plenty of their own first party data. Unlike the traditional TV industry, these digital companies are accustomed to a world where they know exactly what happens with their own audiences and their media consumption; the feedback loop is perfect. These owned metrics are extremely accurate, the sample size is essentially census-level, and they can be super-customized for each specific platform as the needs arise.
This has incentivized these technology companies to create metrics and approaches that preemptively set expectations for advertisers and agencies: there is no other third party to influence or advise on which metrics matter, and these technology companies can effectively self-generate demand for their ad solutions (even without a standardized currency), while developing their own metrics to prove demand.
This all happened because for two main reasons: first, these digital companies happened to be in a position to do so, and second, because there were no good currencies or third-party solutions available. The truth is that digital content consumption has evolved so rapidly, these players make it difficult for external parties to develop standardized ways to measure them, and as a result, vendors in audience measurement have been struggling to keep up. Companies like ComScore, Hitwise, Nielsen, and Netratings were able to develop some basic metrics like page views, but the technologies and methodology they developed cannot keep up with today’s digital, cross-device consumer behaviors.
We have yet to see a currency that consistently fulfills the industry’s needs. At Verto, we’ve built an approach that provides our clients with universal metrics like reach, frequency, and engagement, and includes incremental/overlapping usage metrics into a model which makes TV and digital comparable. Our goal is to provide our customers with insight into things like increases in brand awareness and all combinations of ad exposure on different platforms, within a single-source panel. However, scale is always the challenge, and our metrics still requires more testing and time in the market to get these accepted by the Media Ratings Council (MRC) and the buy side of the market.
Many conference attendees also referred to problems in understanding and/or using digital video metrics. Digital metrics still lack the standardization enjoyed by traditional TV metrics, and there is currently no universally-accepted currency, and or accreditations. For example, Facebook reports “video starts,” while Amazon reports “average minute audience.” The next question is, how can we improve digital video metrics to or establish some form of standardization? It’s a question with no clear and immediate answer, but an important one I’ll be tackling in my next blog post.