Monetizing FAST and OTT with Metadata and User Data
There’s an old maxim in the publishing business that writing is writing but publishing is timing. In the media and entertainment world, the corollary might be that content is king, but monetizing content is a whole other thing. Quality and quantity matter, of course, but every bit as critical as the creation and curation of a content library is the science that supports and sustains the art.
Particularly in the saturated content markets of OTT, FAST, SVOD, and AVOD, delivering great shows is barely half the battle in the struggle to catch and captivate eyeballs and convert viewership into profitability. OTT’s upside, even in the days when it lagged far behind traditional linear TV in audience share, was the ability to target audiences with greater precision, personalize experiences, and leverage recommendation engines to surface the right content for the right viewers.
Of course, nothing makes recommendation engines—and every other means of profit-making leverage—go, go, go like data in its many varieties. Metadata in particular is the key to personalizing viewer experiences and optimizing content portfolios. AI plays an ever-greater role in enhancing the metadata quality and improving recommendations. As analysts focused on the M&E industry and insiders at major media companies playing critical strategic roles agree, data is also essential to understanding customer journeys and making informed decisions on how to reach and retain subscribers and, on the ad-supported side, deliver the right ads to audiences and the right audiences to brands.
Metadata: The Linchpin of Everything?
In a panel at Streaming Media Connect 2023 in November, Chris Pfaff of Chris Pfaff Tech Media described metadata as “the linchpin, the DNA of everything” when it comes to driving ROI for OTT and FAST.
Paul Erickson, founder and principal of Erickson Strategy & Insights, calls metadata “table stakes” for an effective OTT monetization strategy and “a potential force multiplier for overall service ROI, if you leverage it correctly.”
It all comes down to “sustaining engagement,” Erickson explains, “which is fundamental to retention and monetization. It’s directly connected to personalization and search and recommendation. If we think about how efficiently you can target and address specific audiences with your advertising, that specificity is aided tremendously by the metadata that helps advertisers understand the nuances of that audience that they’re trying to hit. There’s a lot of value to be reaped by those parties that can effectively enhance and enrich metadata, typically using AI to help do so, particularly when it comes to equalizing the level of metadata quality and richness in diverse libraries that may consist of new and legacy content from different sources, such as what you might have in a FAST channel.”
Any cost/benefit analysis must also acknowledge not just the return but the investment side of ROI, as well as the opportunities for cost reduction that metadata brings, Erickson continues. “Metadata also helps you understand the specific nuances of portions of your content portfolio that are not performing well and gives you some insight into why, depending on the richness and quality of that metadata. Thus, it can help you optimize your existing content portfolio and inform your future content acquisition strategy.”
Also critical to effective use of metadata, of course, is having metadata that’s accurate and clean. Bethany Atchison, VP of distribution partner management at Vevo, highlights the importance of having well-organized data at a service whose library consists of 800,000 music videos “across a variety of decades across genres” in order to deliver a high-quality user experience. “Organizing our data is really important to keep track of all of that content. We really need to have clean metadata in order to organize everything,” she says. “We have a programming team that uses their expertise to program each of our individual FAST channels, but they need to leverage AI and data to really comb through all of that and surface everything to them because there’s just so much content to go through. For us, it’s really tantamount to the success of our channels.”
Content recommendations on Vevo
Dan Trotta, product manager at Warner Bros. Discovery, underscores the value of metadata by describing how his company and other premium content services use it to intelligently determine which “pause ads” appear on-screen when a viewer pauses a show and where a brand’s ad might be placed and—just as important—where it will not. “Let’s say you’re watching something like Breaking Bad, and you pause on one of those moments in the show where maybe you don’t want somebody walking into the room, and you also don’t want an advertiser’s ad to pop up over that. Streaming services are working toward metadata that runs through the episode to show you when areas are safe and when they are not,” he explains. “It allows you to take this ad format and distribute it across more of your programming hours than you otherwise would have. That’s money in your pocket right there. And it’s a better viewing experience than dropping an ad where it shouldn’t be.”
Pause ads on Max
Contextual vs. Behavioral Data
Much of the advertising that targets specific audiences on the internet—and over the top of it—leverages behavioral data that indicates a user’s interests or demographics to serve relevant user-appropriate ads. Contextual data analyzes the content itself and aligns ad campaigns with interests suggested by that content. Text-driven contextual advertising is relatively straightforward, relying on scans for keywords to match content with relevant ads. The more sophisticated contextual ad solutions applied to OTT content analyze it for sentiment and tone as well.
Eric Berger of Common Sense Networks, which develops ad solutions designed to serve age-appropriate ads alongside children’s content, breaks down the benefits of using contextual data, as well as some of its limitations as currently implemented. “The problem with contextual data today is that people are using metadata that’s created by creators in the case of platforms like YouTube, and the contextual match isn’t that strong,” he says. “The better you can be at finding the actual match between the content that’s on the screen and the advertising, the better it is from the consumer experience, and the higher the engagement is. Sometimes it’s a topic, like a hiking video with an REI ad.
But sometimes it’s also the sentiment and the mood—teamwork, role models, empathy—that connects and draws you in. That’s why many people like contextual more than behavioral.”
Common Sense Networks’ contextual ad targeting and matching
Another advantage of contextual over behavioral data in generating the desired consumer response, Berger says, is its timeliness. “Behavioral is based on past behavior,” he explains. “How many times have you looked for a pair of shoes and then [shoe ads] follow you all over the internet, but you’ve already bought them or made the decision not to buy them? Contextual is much more in the moment.”
Effective contextual advertising is, if anything, even more data-dependent than behavioral, even though it doesn’t leverage user data per se. “It’s really important to get the data piece right and to scale as much of the sea of content as possible. We’ve taken 3 years of content data from evaluating content, and we’ve worked with Deloitte to supersize these datasets using machine learning so that we can effectively catalog the sea of content that is appropriate for kids and be able to slice and dice it through over 300 different types of metrics without any user data whatsoever.”
Part AI, Part Human
As Berger describes, machine learning and AI are critical tools in taking data analysis to the next level, whether for ad alignment or content recommendations. As recommendation engines become more sophisticated and more efficient in their use of metadata and user data to improve recommendations or ad relevance, enhance the user experience, or hold viewer attention longer, the question arises of how much AI factors into the equation and how much human intelligence still contributes directly to the engine’s decision making.
“A lot of what you see on a streaming homepage is personalized to you and your past viewing behavior and what viewers similar to you have viewed in the past as well,” says Warner Bros. Discovery’s Trotta. “The human element there that’s still really vital is pointing the algorithm in the correct direction. That means setting the key input and key output metrics.” The human element also determines answers to key strategic questions: “Are you going to optimize for the most active viewers you can get on your service in a given day? Are you going to optimize for the most viewing hours you can get on your service in a given day? Are you going to optimize for the frequency with which users come back in a month? That gets sped through the pipeline that ultimately becomes what the user sees. But that’s the fundamental question to kick everything off each time.”
“AI is going to be a huge game-changer,” says TVREV co-founder and lead analyst Alan Wolk. He coined the term “FAST” for free ad-supported television during its infancy as an outlet for digitized longtail and legacy content that FAST channel providers were discovering could enjoy a second or third life as commercially viable media, thanks to the improved targeting opportunities available through OTT and streaming. Much of AI’s real power in this area, Wolk contends, is to dig deeper into shows to analyze sentiment and mood, as Berger attests, and make more informed decisions about ad placement and the like.
“There are companies out there like AiBUY and KERV that are actually able to go through and identify what’s going on in different scenes,” Wolk says. “Nobody wants to run their funny commercial during the funeral scene of a show.” Until recently, he explains, “It’s been very hard to identify that. Now they’re able to do that by understanding the emotions involved.”
These capabilities also create opportunities for a level of contextual insight and content-matching unheard of before AI became involved. For example, Wolk says, “If I’m Cadillac, and Matthew McConaughey is in a scene, I want to run my ad there, since he’s in my ads too. It creates a nice synergy. It can even understand what sorts of shows viewers like and personalize their feeds. That’s going to be huge.”
Pfaff argues that the more intelligent, targeted, and curated nature of data- and AI-infused ad strategy might also help to diminish “ad frequency and ad fatigue”—a key concern as streaming monetization becomes ever more ad-centric, with even the most stalwart subscription-based services adding and emphasizing their hybrid tiers.
“Ad overload and poor targeting are always going to be a net negative for engagement or retention, which obviously impacts your ROI over time,” Erickson concurs.
Taking It Personally
With all of the reliance on using personal data that personalization demands, the question arises as to whether there may be a reckoning ahead for the content companies—along with other digital data gatherers—that depend on that data. As consumers become increasingly aware of the privacy they’re giving up—and what their data is worth to the companies that collect it—they might not be so willing to surrender so much information on their preferences and behaviors just to get them to the shows they want a little faster or to train AI models to serve them ads for products they’re likely to buy.
Publishers Clearing House recently did an extensive study on consumer attitudes regarding data collection and usage for a range of purposes, including the personalization of M&E content and advertising. Produced in conjunction with ESHAP CEO Evan Shapiro and others, the findings were published in November 2023 in a white paper called It’s All Personal.
PCH's It's All Personal report
Many respondents seemed not to fully recognize the value of their personal data to the companies that collect it. But they did seem to value the ways personalization benefits them. Fewer than half (45%) said they’d trade less personalized ads for more control over their personal data. The survey revealed an overall passivity about tolerating the types of data gathering that enable the targeting of OTT content and a willingness to give up a little bit of privacy in exchange for the personalization of their entertainment experiences.
“When we looked at people’s attitudes towards advertising, I think we expected people to be a little turned off by personalization of advertising, but that didn’t seem to be the case, especially for younger consumers,” says Shapiro. “They seemed to be totally chill about their data being used to personalize ads towards them.”
“The data was pretty much divided equally among people who want to have personalized content and people who don’t want to have personalized content,” says Publishers Clearing House assistant VP Smriti Sharma. “Everybody talks about privacy, but do people really understand what privacy is? Because clearly the only way you can provide personalized content is by capturing their data and identifying what they like.”
The end, it seems, justifies the means for most media consumers. “If you have the right content,” Sharma says, “people are going to follow it, and they’re ready to pay for it.”
Overall, the Publishers Clearing House study suggests that access to the personal data that powers personalization isn’t likely to diminish any time soon—at least not as a result of consumers acting on their privacy concerns.
Striking the Right Balance
Erickson asserts that as important as personalization, clean metadata, and AI-heightened ad positioning might be, no single factor contributes more significantly to streaming service ROI than sound content budgeting. (This may well be the fate awaiting premium content providers that have overspent on original content and tried to reverse their fortunes with ad tiers, sharing crackdowns, price hikes, and the like.) And naturally, matching the right content to users presupposes having the right content to begin with.
“Content is the lifeblood of any video operation, whether it’s traditional broadcast, a streaming service, a FAST channel, what have you, and it can make or break the economics of your service,” Erickson says. “The right content brings people in, keeps them there, and brings them back. But running out of content leads to them churning or dropping off, and spending too much on that content, no matter how good and relevant it is, will put you upside down ROI-wise. So today, I think we’re finally seeing a serious and judicious review of content spend when it comes to acquisitions and commissions to ensure that overall content budgets are being used efficiently and profitably.”
Erickson suggests that the pause in M&E content creation caused by the writers’ and actors’ strikes of 2023 might actually bring about a bit of a reset that could precipitate the necessary adjustment in content acquisition and development models and a new prioritizing of quality over quantity. “Obviously, great content is always valuable to bring people in and keep them in a subscriber fold from the service perspective. But if you’re not able to commission that great content, we have a situation now where the strike perhaps is allowing the industry to sharpen its skills in a different area,” Erickson says. “Not all great content in your portfolio needs to be original commissions. You can license and leverage some amazing content from around the world if you internationalize it properly or it’s already internationalized. And as your audience base over time gets more diverse, and retaining them as a business gets more important, licensing gives you a lot of agility to make sure you are serving or super-serving all of those particular nuances of those unique audience segments. Whether you are putting content in the portfolio for an AVOD service or SVOD service or FAST channel, you still need to have some level of content agility, but that content spend can shift from being prioritized less on the original commissions end and more towards very prudent licensing of content that’s going to serve those particular audience segments.”
In the absence of subscription revenue that underwrites SVOD services, what's the ideal balance of original and licensed programming for FAST brands, and how does it differ for smaller and larger providers? ESHAP's Evan Shapiro, TasteMade's Evan Bregman, and Chicken Soup for the Soul's Philippe Guelton dissect the intricacies of FAST monetization-particularly for niche outfits-in this clip from Streaming Media Connect 2023.
Personalized data is critical to video content monetization, whether in a subscription-based environment or an ad-driven FAST channel. Coming at the topic from both angles, Starz' Rob Collins and Revry's LaShawn McGhee discuss their organizations' strategies for leveraging personalized data in this clip from their panel at Streaming Media Connect 2023.
Quickplay's Naveen Narayan and Reality Software's Nadine Krefetz discuss how gathering the correct data can improve a streaming platform's understanding and knowledge of its audience, yielding enhanced user engagement and content monetization, in this talk from Streaming Media West 2022.
How much can the personalization of advertising experiences increase user engagement and ultimately help to reduce churn from ad-supported streaming tiers? Evan Shapiro of ESHAP, Darcy Lorincz of Barrett-Jackson Auction Company, and Dina Ibrahim of San Francisco State University talk about the rise of deeply immersive catered ad experiences and the increase in monetization ownership in the era of services like Patreon and Discord.