• CTV Content Metadata and the Need for a Standardized Taxonomy

    As TV and digital advertising converge, it’s become even more evident how different the ways of buying and selling media are in each ecosystem. This reality has created some key challenges for both brands and media owners who seek to operate across platforms, and these challenges will only inhibit the free flow of money if they persist.

    Advertisers generally want to deliver targeted impressions across a mix of programming, irrespective of whether that content is delivered on a set-top box or an IP-connected device. However, each of these environments offer vastly different capabilities and operates on different protocols. The resulting asymmetry, as you might imagine, often leads to frustrations and hurdles.

    Measurement and identity are two pronounced examples of these challenges that are garnering headlines today. An often overlooked source of friction, however, lies in the lack of standardized metadata when it comes to transacting on signals like content; specifically, the manner in which program-level information is structured and shared among media owners and advertisers.

    Programmers have premium connected TV and OTT inventory that they are looking to monetize, often through programmatic channels. In making their inventory available to buyers, they typically relay some key programming information at their own discretion, through a bid request. That request will include details on the content in which the impression opportunity is available; for example:

    • Series: Guy’s Grocery Games
    • Producer: Food Network
    • Genre: Food & Cooking
    • Season: 18
    • Episode: 9


    These signals offer pre-bid transparency provide immense value to ad buyers on the other side, who can now understand exactly what they're bidding on for purchase. You could equate this to the window sticker on a new car.

    Within traditional linear TV environments, agreed-upon operating standards for program metadata, such as EIDR, have been in place for years, though they too have had their own challenges. In digital, the Interactive Advertising Bureau (IAB) has taken steps to provide a specification for attribute names when exchanging this type of information in a programmatic fashion, but there is still no unified taxonomy for standardizing the format or content of what’s passed within those attributes.

    Specifically, the IAB open RTB Content Object includes basic parameters for configuring bid requests to detail more than 30 different content signals. These include genre, series, title, language, content rating and even production quality, among other things. The challenge, though, is that there is no set standard for how to input and structure said information.

    As an example, for a live sports event where impressions are available, one publisher might pass the opportunity through as “Sports'', one might note it as “Football” or even “American Football,” and one might simply refer to it as “NFL.” This, bear in mind, is assuming no spelling or spacing errors are made!

    While some publishers make use of the IAB specification, not enough do. There are a number of reasons for this; a big one is that a chicken-and-egg situation has developed between buyers and sellers. Buyers often can’t transact on metadata at scale because it’s too labor-intensive to decipher, and it’s largely unreadable. Sellers are reluctant to pass the information because buyers aren’t actively transacting on it, and it takes quite a bit of effort to maintain for programmer catalogs.

    It’s a missed opportunity for buyers to gain greater pre-bid transparency, and for publishers to earn premiums on their inventory.

    So, how do we solve the issue at hand?

    First and foremost, moving towards a uniform taxonomy — a standard way of structuring, defining, and inputting key content information — will make it much easier for sellers to configure their request and buyers to respond in kind. There is a lot of progress being made in this regard, and it’s something we’re optimistic will continue.  

    Second, normalization is key. Taking raw inputs and normalizing them into standardized signals alleviates a lot of the burden placed on publishers and brands.

    Third, agencies and buyers should be asking for more metadata signals from their media partners in programmatic (many already are).  Their continued advocacy will only help to further move the needle. And when programmers provide it, they should charge a premium to make sure they aren’t getting their best pieces of content cherry-picked.  

    There’s work to be done, but the value of metadata is unquestioned. Metadata will play an increasingly critical role in the way connected TV media (and TV and digital media more broadly) is transacted, especially as ecosystems continue to converge. Arriving at a standard taxonomy is key to making this a reality.