Getting Personal

You are starting a new OTT service. You have a brilliant recommendation engine supplier lined up that will drive viewing suggestions on an individual basis to every one of your brand-new consumer base. Exciting. But also dangerous.

Broadcast channels have a challenge of taking an audience they know little about to a customer base they can build a close analytical relationship with.

All recommendation engines work off algorithms (simple or complex) designed to tailor suggestions based on each consumers’ actions, previous viewing, and other data points. But therein lies the first problem; you don’t have any real data on your consumers until you start.

The deeply mined anonymised data banks of your chosen recommendation supplier are designed to make up for this, but what was the source of that data? Does it understand the fine distinction of cultural or genre nuances, apply equally well to a generic film service (e.g. a Netflix) and a niche OTT service (e.g. a cooking channel).

I’m not questioning many of these facility’s ability to learn, but I do question the wisdom of launching without data that applies to your own unique OTT offering. Prudence and cost savings may yet back an approach that starts with curation and moves to personalisation when you really know your customer.

Visit to find out more.