RTÉ is Ireland’s national media organization and Ireland’s largest media outlet for both TV and online publishing. As a public service media provider, RTÉ has to deliver a wide range of content that appeals to all tastes and audiences.
However, even with the great variety of content delivered by their newsroom, their reader base was heavily skewed towards Breaking News and Sports sections. Since their article recommendation were based on publication date and page views, more articles from those two sections were being displayed, increasing one specific audience and undeserving others.
With Readersight, RTÉ content was explored by the reader base as a whole, increasing audience diversity and encouraging the reader to discover a wider variety of content across more sections.
“Working with the Readersight team has allowed us to balance the algorithm to ensure that we are promoting content that is engaging and making an impact but also surfacing other more niche content that normally might be missed"
Head of RTÉ.ie
Serve audiences that have been underserved, especially younger female
Move readers across all sections on the website
Display a greater diversity of content to readers
Increase the loyalty across a broad spectrum of content
Improve CTR of recommenders
Tailored Recommendations Algorithm
We developed and tested 14 different variances of recommendation algorithms to find out which one works best to match RTÉ’s objectives
Same site, but smarter. We built a tailored UI for the “Read Next” & “More for you” sections to match the existing website style on desktop and mobile.
Our journey: from implementation to results
Implementing Readersight is simple as ABC, all the work was on us to build a tailored UI to match RTÉ’s website and calibrate the recommender. By receiving data in real-time our AI algorithms learn using a foundation grounded in metadata. This metadata continued to develop over a couple of weeks, ultimately leading to learning the most successful content to each individual. Teaming up with RTÉ editorial team, we added weighted bias to the sections and articles focused on young female preferences and behaviour.
In total, we tested 14 algorithm variations and 2 main winners emerged; The former increased CTR by 32%, however, was keeping the polarity of the audience on Breaking News and Sports sections, undeserving other categories. The latter increased CTR by 30% and significantly improved the diversity of recommended articles. In this algorithm, the content was explored by the reader base as a whole, encouraging readers to read other sections and topics. Delivering tailored recommendation and presenting content variety, encouraged readers to consume more articles, grew new audiences and increased loyalty
“By understanding the behaviours and preferences of our audience in real-time, we can present the "next best option" to the reader. This had led to a deeper and broader engagement with our content across the site.”
Head of RTÉ.ie