Skip to main content
Industry Contributor 10 Jun 2019 - 2 min read

Machine learning halves unsubscribes at News

By Paul McIntyre - Executive Editor

News UK says a self-learning algorithm dubbed ‘James’ has helped halve unsubscribes (Digiday).

 

Key points

  • News using machine-learning to serve subscribers right content at right time via newsletters across The Times and Sunday Times
  • Trial used 300,000 subscribers and registered users versus control group. Results showed 49% fewer unsubscribes
  • News plans to roll algorithm into data management platform, use for editorial commissioning, and white label to other publishers
  • Project originally funded with €1m from Google's Digital News Initiative

In a world increasingly dominated by social media, News is taking Google funding and applying AI to newsletters. And it seems to be working, in the UK at least. If it can help deliver personalisation at scale and move from preventing unsubscribes to driving paying subscribers, the publisher's claims of "making a CRM nirvana" will be justified. 

News insists that its stories remain human created and curated, with robots only improving delivery.  But other programmes funded under the same Google competition include robots writing the news. 

News agency Radar - Reporters and Data and Robots - is a JV between Press Association and start-up Urbs Media. It launched two years ago with a €700k Google grant. UK media firm JPI, which manages around 170 mostly regional news titles, takes around 700 robot-driven stories a week from Radar, according to the Financial Times.

Which sounds suspiciously like CFO nirvana. 

What do you think?

Search Mi3 Articles