JAMES, a Digital Butler, the AI for personalising news reading triggers, developed in partnership by Twipe and News UK has won silver in the Global Media Awards for ‘best new digital subscription initiative.’
Based on a one year project with The Times and The Sunday Times in London, JAMES has been developed by Leuven-based media tech startup, Twipe. The goal of JAMES is to serve newspaper subscribers the kind of content most relevant to them based on their reading patterns, at the time and frequency most suited to their habits, using a self-learning algorithm. JAMES gets its name from the fact that it acts as a digital butler, serving up content to individuals in the way and format that best suits them. Just as a butler does, JAMES learns and knows the preferences of the user and discreetly serves the right services to them as individuals, rather than at the segment level, long considered best practice for customer marketing.
“Just like an ideal butler, JAMES observes you, remembers where, when and how you like your news to be served and then does it for you without you even noticing." – Danny Lein, Twipe CEO and Founder
A team of over 20 people, based in Leuven, Spain, the UK, and India worked on a variety of experiments over the one year project which saw more than 14 million emails sent. The project also benefited from an investment from Google through it’s Digital News Initiative which funds news innovation projects.
What made the experimentation with JAMES different than other personalisation research is its unique focus on news content. Outside of the news industry, there’s been a lot of great work on personalisation, such as the strong recommendation engines from Netflix and Spotify. However, personalisation for news can still be difficult, as news is different every day and inherently different than music, movies, or a product from a catalogue. It is important also to remember that people still come to newspapers for news, so publishers must find the right balance between personally interesting stories and popular stories. With the hybrid model built in JAMES over the course of the one-year research project, we are able to develop a proprietary algorithm that mixes popular stories and stories that are personally relevant. This resulted at The Times in a 49% reduction of subscription cancellations for the analysed cohorts.
Twipe is now on a journey to productize the technology for the broader media sector and is already working with launch partners in the UK and the Netherlands.