Taming the Many-Headed Problem of Video Content Discovery

Let’s face it: The OTT landscape is becoming increasingly fragmented. The promise was that consumers could create skinny bundles of content they wanted (from across providers) in order to replace endless channels in a traditional cable operator’s electronic programming guide (EPG). The reality has become a siloed monster, as brands split off from aggregator services (like Hulu and Netflix; remember, Hulu was formed by Fox and Comcast) into their own direct-to-consumer offerings. It seems there’s an announcement of a new OTT offering every day. If consumers want shows from all of them, they must now have a multitude of OTT subscriptions.

The underlying issue to this isn’t the fragmentation itself, though. It’s that consumers see all of their OTT services as a single “TV experience.” Growing up with the EPG (and, to date myself, TV Guide), consumers want to access all their content in one interface, regardless of its location. Thankfully, some OTT platforms, like Xfinity and Liberty Global to name a couple, are busy integrating providers like Netflix into their programming guides so that viewers don’t have to fumble around with a bunch of apps. (Oh yeah, and which app do they use? The one on the smart TV, or maybe on Apple TV, or how about Roku, and don’t forget the Amazon Fire TV Stick…). Other innovative companies, like Caavo, are attempting to solve the problem in different ways—by acting as a proxy to all of a viewer’s subscriptions and aggregating the content into a single view. (Apple does something similar.)

But that only partially solves the problem, because viewers are still left with considerable angst about what content is on what provider and, more importantly, what content they might enjoy watching. As OTT providers have incorporated recommendation technologies into their user experience, consumers have become used to them. The idea of returning to a dull, old EPG offered by the cable provider that doesn’t supply any sort of recommendation seems barbaric. But how can viewers get recommendations across all their services? The short answer is, they can’t.

The first reason they can’t is logistical. These OTT providers are all very siloed. They don’t want to share information with providers they see as their competitors. Why would they want the viewer to get recommendations for content that’s in a rival’s service (even if the consumer doesn’t see it as a rival)? These services have become walled gardens despite the open web technologies on which they are often built. I can guarantee you that most of them have RESTful APIs, can spit out JSON or XML, and have a metadata format they could share.

And that is the second problem right there: metadata. Unfortunately, most OTT providers and video distributors roll their own metadata framework that describes their video assets. Organizations like MovieLabs and EIDR have implemented ways to hook together video content data sources (the MovieLabs ontology is particularly interesting), but that doesn’t solve the metadata issue. There is, simply put, no standard way for the streaming video industry to describe and represent a video asset programmatically. And that creates a major problem for content recommendation and discovery. If none of the services are “speaking the same language,” it becomes very hard to recommend content to the viewer, let alone to search across the services.

In order to unify this new TV experience, the industry must come together and agree on a standard data structure to describe content. Then, OTT services must be willing to open their gates so that description data can be shared and consumed elsewhere. Everywhere. The thing they fear right now could be their greatest asset: Instead of worrying about recommending content on a rival’s platform, perhaps they should look at the opportunity for viewers to be recommended content that’s on their service? Obviously, it goes both ways.

[This article appears in the July/August 2019 issue of Streaming Media Magazine as “The Growing Problem With Content Discovery.”]