Data is remaking media, and understanding this allows companies to shape growth opportunities across many fronts. I‘ve approached these changes by focusing on efforts that give the best return on monetizing content. After ten years in the data business, here’s my perspective on the latest shift in media.
We’re not going to stop watching TV, consuming video, or checking Facebook anytime soon. As these practices endure, media companies must continue to support them, while also growing greater value from their digital assets. These major forces in our time—media consumption and data’s insane growth—show dramatic momentum for actionable business, and to stay relevant, media companies are embracing data-driven strategies more than ever before.
From “anywhere TV” to “going digital," each incarnation has challenged the way media is consumed and monetized. Today, media is no longer just about content and programming; it’s about maximizing data’s role at every point in the cycle. Staking out the entire media landscape is not possible, but from looking at three traditional areas of linear TV, advertising and sentiment, and IP and file-based workflow, we can see capacities of great change. I was a part of building these systems. Now it’s time to see how data is influencing their future.
Let’s situate our perspective with a top-level look at data and media trends. This moment in time has recently been referred to as the Reinvention of Media, evolving at an unprecedented rate. Consider:
It’s difficult to totally conceptualize these figures and trends, let alone leverage them. So, just how can they be actionable in the world of traditional media investment?
As major stakeholders assess their best strategy, they know that media’s best friend is data. Data provides deep value to content itself, in both the context and measurable metrics. Today, data impacts everyone in the media community, from creators of images and video, to packagers and distributors of content, and advertisers monetizing consumption. Data is driving content monetization because it provides vital information along the value chain. At capture and post, creators set dates, duration, file type, location, director, genre, copyright, and much more. The expedition continues as digital asset management systems (appropriately called “DAMs,” managing the tide of media) feed distribution, where data architectures manage linear TV, Over-The-Top (OTT), online videos, and social networks. Finally, data taps into online consumption that matches content to people’s valuable contacts, location, and behaviors.
A parallel universe is also moving with data in the form of metadata about customer behaviors—collected, managed, and analyzed with images and video. There’s so much value now. Metadata is also driving collaboration on cloud, asset creation on encoding platforms, and management of inventory for digital asset managers. This all represents found money because these data-driven processes contribute to a feedback loop to give greater value to the asset.
Applications even ask the question: “Will you ALLOW me to use YOUR data?” The answer from billions of consumers: “Yes, a thousand times yes,” often without a precise understanding of the monetary value of their agreement. We’ve moved far beyond impressions for ad value. Every time someone switches on, watches, clicks, or shares, they create more data and viral value. The feedback to the provider has its own special metric, in the form of an algorithm, and now, content accompanies the very data that drives advertising.
True, linear TV is declining, especially among young people. But one major report shows they’re still watching 20 hours per week, and older people (50+) are at over 40 hours. With billions of dollars at stake for millions of consumers watching TV, the linear TV market is still very relevant. In fact, 50% of media spend is for linear TV, and data and analytics are helping to maximize that revenue.
Today, a major strategy for linear TV is to secure upfront sales (say, 60% of the programming inventory) and then sell to the scatter market (what’s left) or filler (e.g. late night infomercials) to complete the total buy. There’s a built-in liability to manage as well. For ad sales, it’s all about deal making, and media salesmen need data and insights to prove more value to buyers. Scheduling and protecting this media inventory sell requires a ton of data on consumer behavior, traditionally dominated by Nielsen.
At a recent event sponsored by Georgia Tech’s Business Analytics Center, the highlight was Wes Char’s presentation on Turner’s activities to maximize viewership, ad revenue, and cross platform strategies—all with better data, insights, and analytics. As SVP, Analytics, Data & Decision Sciences at Turner, Wes leads a growing team that’s assessing the value of the programming inventory of key properties from CNN to TBS to Cartoon Network. He’s leveraging new digital strategies from ideas coming from marketing experts around the world, such as i-Com, the quickly growing Global Forum for Marketing Data and Measurement.
PayTV (in forms from cable, IPTV, DTH, and virtual MVPD) represent a much more difficult challenge. The data collected and analyzed from the home set top box (STB) back to the cable headend makes targeted advertising quizzical because providers don’t manage it the same way—it’s proprietary. Even though great understanding could be gleaned from the combination of audience consensus with actual second-by-second behavior for targeted ads, impact for greater sales revenue would still have to be shown. That’s many data sets in play and a host of analytics. But, watch this space. With 100 million subscribers and 300 million set top boxes, the top MSO’s see this as a major initiative to use data to drive sales and decrease churn or cord-cutters. Not to mention that MSO’s will often use STB generated viewership data in licensing fees negotiations with a programmer.
What data should be looked at in media? The short answer: all of it. While programming still has a cost basis, a shelf life, and target audiences and communities, it is changing due to the collection of data and ratings. Ad strategies can better understand which communities are getting the message and how many times they saw it. The audience measurement has changed as well. Beyond the living room measurement alone is the combination of TV with a tablet, laptop, or mobile device. Viewers just don’t watch a program solo anymore. Take NCAA March Madness as an example. People watch the live broadcast, but they’re also tweeting, finding clips of past games, and reading player stats simultaneously.
Big news. Fox, Viacom, and Turner are launching OpenAP, an effort for advertisers to employ the same strategies used to measure digital media consumption. The program moves beyond demographics and into communities. And it’s driven by data. New analytics apart from age and gender are seeing massive investments. Now, what someone does or who they associate with as an active “community” takes the lead. For example, the hiking community embraces all ages instead of looking like a classic 25-49 demographic. How will data provide new understanding on what’s important to hikers? It might not just be good boots or down jackets. Moving from living room metrics to activity-based metrics is likely a better bridge from TV to online and mobile programming value.
Abel Tamayo of Envision Digital Media, a thought leader for technology implementation in media remarked, “Programmers are well aware they are not getting all the value from content produced for TV. An important pathway will be to embrace this sense of community, not demographics. Data can enable that.” Building on this concept, data can also monetize long tail content and older assets that may not have big viewership on linear TV, but may find a niche audience if presented in the right way on the right platform.
Sentiment will also be a major factor in program value. If reactions to content or a particular advertisement can be measured, a demand side platform (DSP) for video can be crafted as well. Oscar Wall, a senior director at Ooyala said it well: “Video analytics will grow to be about measuring reactions in close-to-real-time and making changes to content or ads immediately as needed. The greater impact of this will be seen as engagement analytics begin to inform content production (TV, movie, news) curation and/or commissioning. Time-coded metadata can be collected by TV writers, producers and directors to tie social media comments back to an exact moment or event on a broadcast. These audience reactions can then be used to inform future storyline strategy.”
Semantic intelligence is also a growing strategy for major media companies leveraging published content. For a long time, digital publishers have used metadata for online assets and images to create a digital demand and supply for customer behaviors. Now, deep analytics, fueled by unstructured social media and online data, is driving a field of semantic advertising. The centerpiece of sentiment is what someone feels, and measuring the ontologies of content, relationships, and behavior is now very possible—all with a data strategy.
Today, the highest value advertising target might well be in semantics, but a progressive data approach, beyond older relationship database structures, will be needed. Semantics in the data world simply provides understanding from something referred to as “triples,” containing the subject, predicate, and object, all just data terms from a NoSQL model. It’s a different data model, but may be able to handle bigger data even better. Look to companies such as MarkLogic for its leading NoSQL approach or to Expert System for sentiment integration ontologies and text analytics. They show how transforming human behavior into useable data or establishing causal relationships for sets of metadata can set up the ultimate analysis to understand just what people want to see when they want to see it.
In short, everything digital users do, watch, and buy can be tracked for their sentiment and their habits. Now, there is a feedback loop back to the provider on both the content and the consumer behavior, and data architectures continue to grow with great toolsets on cloud, data integration, and Kafka, to name a few.
Finally, let’s consider the very people that built the media business from the ground up, the broadcasting and IT staff in media. This workforce runs the vast media infrastructure and is challenged on how data is changing the operational view. As the broadcast industry moved from analog to digital SDI to an IP file-based workflow that fed many platforms, it has tested the patience of technical leaders for just how much is needed and why. Media strategies must encompass rights, distribution, and management across platforms. That’s where data strategies can help.
IP and file-based workflow is fueled by the data and metadata transported over networks between applications and storage. There are standards, practices, and rules. But the growth is beyond just managing the metadata. Management of the asset inventory has to be translated to monetized content value, and is now done with an assessment of key analytic trends.
For cross platform strategies in advertising, TV programs can drive online viewership and vice versa. To do that, asset archives (metadata from a variety of archived inventory) must be combined with other types of data sets (e.g. social network data, online consumer data), and finally fed into a data lake. This is mostly a data warehouse in a new form, built from a collection of raw data and current data assets. If managed correctly, this new approach can be powerful and drive an analytical view or visualization to help make better decisions on the value of programming. But it will take time.
One of the top experts in the field, Gary Olson of GHO Group, suggests that broadcast engineers need to look more closely to the IT industry for new ideas and concepts. Data science is one of the practices being used to monetize and optimize the use of big data, so why not capitalize on it for media?
Data is all about the value of the asset managed in a dynamic workflow. It’s not that important to completely manage all the zettabytes or find a data-mined piece of customer behavior lost in a digitized stack. Embracing how data scientists think—working up from the data, not down from the application—is an effective approach. Once that’s done, levels of data value in storage, cloud, premise, or hybrid can be established. Then, media data lakes, which can be a good approach to analytics, and metadata can be correctly implemented. This is central because metadata reveals all, providing context, metrics, and semantics to the content.
What keeps media companies relevant is the same thing that keeps me engaged after all these years—realizing that the rubric for value in media is a moving target. What counts now is how people are generating and consuming data, with video taking a step back to be in partnership, not preeminence, with data systems. New data really is remaking media, across linear TV, advertising, and workflow, and will be a cornerstone for the next ten years and beyond.