UTELLY tell us about their aspirations for Data Pitch and beyond.

Television has evolved greatly in recent years, and consumers are faced with more channels, content and interactive services than ever before. The amount of choices available is becoming overwhelming for TV watchers, and without an easy and effective way for consumers to find the TV shows they like, the abundance of content could become a disadvantage rather than an advantage for TV providers.

UK-based startup UTELLY are on a mission to reinvent content discovery for today’s TV. UTELLY’s platform aggregates data from multiple sources, and uses advanced Machine Learning (ML) techniques to build content suggestions that ‘learn’ from user behaviour. We spoke to CEO Romain Eude about their plans for Data Pitch.

What do you hope to achieve on the Data Pitch accelerator?

We hope to improve our set-top box content recommendation engine and create close ties with a leading European TV operator.

What shared data will you work with and how will you use it?

We are working on the Personalised Entertainment Challenge (developing the next generation of multidimensional recommendations) with Altice Labs Portugal, who are providing us with anonymised set-top box usage data. We will enrich the usage data with our programme metadata, and then build ML models to help recommend content to users based on both their historical and current viewing habits.

Why do you think it is important for startups to work with large scale data providers?

It is easy to build solutions that work at a small scale, in so-called “lab conditions”, only to see them crack when attempting to scale in the real world. In the TV world, everything is about scale, so working with a substantial dataset, such as the one we’re accessing through Data Pitch, will help us to develop a better solution for our client’s needs.

What’s the best thing about working with data?

Data trumps assumption – it keeps you focused on what’s real!

If you could change one thing about the data ecosystem what would it be?

We would like to see data anonymisation processes agreed within each industry in order to foster more data sharing and trust between companies.