What we've done

The BBC is unique in its effort to not only create great content online, but also contribute to the greater web ecosystem by linking to relevant, quality content outside of our web sites.

In the 2008 service review, the BBC Trust mandated the BBC to improve its external linking. This sentiment was reiterated in the 2010 Putting Quality First strategy, which set a target to double the traffic sent to external websites from by 2013.

We collaborated with the team behind BBC News Online to research ways to help them reach that goal and developed a prototype delivering semi-automated recommendations of articles from a selected list of news sites.

Why it matters

Linking to content outside of one's own playground is crucial to the fabric of the web itself, and there is an element of the BBC aiming to “be a good web citizen” in our approach to external linking.

External linking also benefits the BBC in general, and BBC News in particular. Links to specialist content, to different takes on a story or to the sources of our reporting not only enrich our audience's experience, they also help secure our trust relationship with them.

Our goals

Our research and prototyping work aimed to explore three questions:

  • What is the current culture and practice of external linking in BBC News Online?
  • Which tools and methods could we use to improve the effectiveness of our external links from the online News website?
  • What can the role of social media be in our external linking practice?

How it works

In the first phase of the project, we researched the state of the art in external linking, drawing from the expertise of related fields such as that of online advertising.

Following a review of research literature on the subject which made it clear that a strong external linking policy with tens of thousands of links to other publications was extremely rare, especially in the News world, we set out to understand where we could improve our linking efficiency, and pursued two complementary approaches.

The first was to examine data: given statistics on page impressions, number of links and the number of times an individual link was clicked we were able to analyse the efficiency of links and categories of pages, as well as produce a number of visualisations. The second approach involved conducting interviews with three BBC News journalists to discover more about how they link, why they link and the problems they face.

Our study helped us build several models of how our linking practice, both automated and manual, could be improved to reach the goal of doubling the external traffic driven from BBC News Online to other publications.


By combining the models developed during the first phase of the project with the audience insights made available to us, we built an internal prototype called “What The Papers Say” (unrelated to the eponymous radio programme) exploring:

  • How, given a set of BBC news stories, algorithms could pick relevant related stories from a set of news feeds;
  • How picking related stories from a specialised set of news feeds (e.g politics or technology) would yield significantly better results than recommendations based solely on similarity to our story;
  • How links from the twitter feeds of our correspondents could be used to surface interesting links related to the news
  • How a very lightweight editorial interface could help improve the quality of recommendations from the initial algorithm-based selection


People & Partners

Project Team