The smart money is on personalized recommendations for live TV.
One of the most prevalent perceptions among TV subscribers is that there’s “nothing on to watch.” In truth, there are so many choices available that it’s virtually impossible for consumers to discover what’s really there. The challenge is proactively presenting subscribers with personalized content recommendations, instead of forcing them to work hard to find something they like to watch.
Recently, Cox Communications became the first MSO in the U.S. to launch a new class of personalized TV recommendations as part of its groundbreaking Trio program guide. The guide presents recommendations across live TV and VOD content choices that can be personalized to individualized members of a household, and it’s a bellwether of what’s to come to TV viewing across the U.S.
Cord-shaving debate aside, many consumers are actively questioning the value of their subscriptions given that the average subscriber watches only a handful of channels from the hundreds for which they pay. It’s a real problem operators must address immediately, and the smart money is on personalized recommendations.
Search is only useful if the consumer already knows exactly what they want to watch, which is only about 4 percent of the time. The key is to add intelligence. Intelligent recommendations rely on a combination of robust metadata and a variety of recommendations algorithms, on implicit and explicit data points and direct user feedback via likes and dislikes, and on consumer viewing behavior. With new kinds of personalized recommendations and intelligent navigation systems, operators now have the key to unlock the value of their content catalogs and demonstrate the full value of their subscription packages.
Most U.S. viewers, however, only have experience with rudimentary video recommendations. Take Netflix, for example. In households with multiple viewers, such as a family with kids, receiving recommendations to watch “Sons of Anarchy” alongside episodes of Scooby Doo is a common type of problem. That’s why individualized recommendations, like the kind being deployed by Cox, make sense. Even without an individualized login, recommendations engines should take into account factors like time of day, or days of the week, before presenting recommendations. The fact that the set-top box is a multi-person device requires a more sophisticated approach than simply delivering generic household-level recommendations.
Other rudimentary recommendations systems rely solely on “collaborative filtering” techniques (i.e., “people who watch this watch that”). A major limitation here is the tendency to surface only popular content, and not more obscure choices that would actually be more relevant or enjoyable to the viewer. More importantly, from a business perspective, it doesn’t showcase the full value of an operator’s catalog. In addition, brand new shows, or shows that have yet to air, wouldn’t be flagged in a system like this. That’s why collaborative filtering should only be a part of a more comprehensive approach to recommendations.
With so much attention on catch-up TV, VOD and DVR viewing, it’s easy to forget that the majority of consumers actually start with their EPG and watch live, linear TV a majority of the time. That’s why TV operators need to support intelligent navigation and personalized recommendations in the live, linear EPG.
Providing recommendations for a few thousand mostly static VOD choices isn’t very difficult. By contrast, delivering real-time recommendations for dynamic, linear TV and then personalizing it for an individual viewer is a considerably larger challenge. A typical 500-channel EPG offers more than 200,000 TV choices in a two-week period. This is why making subscribers rely on search alone is a losing proposition. A typical linear EPG catalog is updated every 15 minutes, meaning that any recommendations engine and intelligent navigation system needs to work at a much greater scale than those designed for a simpler OTT-like VOD environment. Operators such as Cox and Liberty Global are at the forefront of providing live, linear TV recommendations in addition to VOD as part of a new and better TV experience.
Research and results from large-scale deployments have proven that delivering better TV recommendations can have a profound impact on business results. Fundamentally, any recommendations deployment needs to support the service provider’s current business objectives and have the inherent flexibility and scalability necessary for the system to evolve. For example, one operator might want to set parameters that show viewers recommendations options within their existing subscription package only in order to emphasize its value. Another operator might want to show 80 percent of the recommendations options from its existing package and 20 percent from a premium tier as a means to upsell.
The technology now exists to let operators configure the parameters that best meet their business objectives, and forward-looking operators are building them into their new TV services. Personalized recommendations can provide a powerful tool to reduce churn, market new or underexposed content, increase on-demand sales, and improve consumer satisfaction.
Email: pdocherty@thinkanalytics.com
In the next issue, Clearleap chief architect Jim Tanner will tackle three of the biggest pain points for providers that seek to offer multi-screen services.
