Fighting for quality news media in the digital age.

  1. News
July 19, 2022updated 07 Oct 2022 7:10am

Reach creates AI-powered content recommendation tool for articles

By Alexandra Turner

The UK’s biggest commercial news publisher Reach has created an AI-powered audience engagement tool aiming to provide “relevant content recommendations” to readers.

The AI tool, Neptune Recommender, aims to increase page views, user dwell time and overall engagement.

Widgets will be shown on news sites within the group that contain relevant content suggestions from other Reach brands. Topics are kept current, as the tool updates each time the user visits a page.

Since the start of the year, page views through Neptune Recommender account for 40% of Reach’s page-view growth, the publisher claimed.

The tool is now being woven into every Reach site, meaning more than 42 million monthly readers are offered more personalised content choices than before.

This is the latest addition to a number of digital tools that Reach have launched in preparation for a “cookieless future”, including Mantis, an AI-powered contextual safety tool that can detect whether a story is safe while not delaying publication. Google has said it will phase out third-party cookies on the Chrome browser by the end of 2023 and publishers have been searching for ways in which they could more efficiently gather first-party data and then use it more effectively.

Reach chief revenue officer Piers North said: “With a cookieless future on the horizon and more demand from ever from advertisers for contextual tools, we’ve used our unique scale to our advantage… Neptune Recommender allows us to really dig down into our data to match up our journalism to individual users, strengthening our loyal audiences across our networks, and keep them coming back for more.”

Neptune Recommender combines the AI behind Mantis with Reach’s first-party data, including the information given by users when they register by, for example, logging in or signing up for an email newsletter.

It uses machine learning based on category, concept, engagement and sentiment which, Reach said, leads to higher accuracy than using keywords. Suggestions based on positive sentiment are then shown in a widget.

The tool works to recognise not only the category of previously browsed material, but also the concept which has attracted the user, meaning personally relevant and helpful material can be served up to readers.

Terry Hornsby, Reach group digital and innovation director said: “This is the future; personal, accurate and relevant content recommendation that feels tailored for you, honing in on what you are interested in at that particular time and what you will actually find helpful.”

He added: “…we need to recognise and treat our users as individuals, and we’ve focused on the technology and data strategy that can step up to the plate to do that.”

Reach owns more than 130 national, regional and local news brands, with a monthly reach of 48 million readers across print and digital.

Neptune Recommender will be made commercially available to other publishers in the marketplace, and is customisable and adaptable to advertising.

Press Gazette is hosting the Future of Media Technology Conference. For more information, visit NSMG.live

Picture: Daily Mirror screenshot

Topics in this article : , ,

Email pged@pressgazette.co.uk to point out mistakes, provide story tips or send in a letter for publication on our "Letters Page" blog

Websites in our network