Get a demo

Why recommendation engines are overrated

One of the holy grails of the VOD industry is to provide great tools for search and discovery of content. The use case of being able to provide the right selection of content for each consumer exactly when they want it is very attractive. This should lead to higher consumption, higher consumer satisfaction and ultimately higher revenues.

However, everyone who has ever used a VOD service knows about the frustration that comes with recommendations provided by the service. Since algorithms are based on usage, ratings of large customer audiences, recommendations tend to gravitate towards content recommendations, which are “obvious”. In itself, such recommendations will drive usage for that main stream content, further pushing it higher in the rankings. Too much reliance in a user experience on recommendation engines can give a static content selection for a large part of the user base.

Don’t get me wrong, this content presentation still provides a lot of value to consumers. I’m just saying that the same or better result could easily have been presented with editorial recommendations. I think there is an over reliance on programmatic recommendations on the market, and a perception that a programmatic recommendation engine is a “must have feature” for any VOD service.

Most recommendation engine costs are driven by how much they are used. It’s clear that for most VOD services, there is a good business case for only using recommendations when needed, complemented with simpler business logic and editorial recommendations for a larger part of the audience. For some VOD services, editorial recommendations and promotions may be a better choice than using a programmatic recommendation engine at all.

CUSTOMERS