From a speech by Peter Horrocks, head of the Beeb's TV News Services, to the Reuters journalism institute at Oxford (posted at the BBC's The Editors blog):

Some excerpts from the Nov. 30 post:

But as well as knowing how much people are watching and what type of people are watching, we now have much more detailed information on viewers' responses to individual news stories. It's this change which is fundamentally altering the journalism business. BBC News online has realtime statistics for its most-read stories. On a big news day, a terror alert or a war in Lebanon will get most hits. But on a quiet day a more off-beat story is often at the top. For instance, who would have thought the story about the Sudanese man forced to marry a goat, published in February, would still be attracting enormous traffic in September. (And who could resist the story: "The goat's owner, Mr Alifi, said he surprised the man with his goat and took him to a council of elders. They ordered the man to pay a dowry of 15,000 dinars. 'We have given him the goat and as far as we know they are still together', Mr Alifi said") Or the fascination for the 13ft python which burst after eating a 6ft alligator. (Great headline: "Snake bursts after gobbling 'gator")

And the online audience is often highly interested in stories of entertainers and celebrities. News of Britney Spears' new baby attracted 350,000 readers in one day and the interest in the honeymoon of Tom and Katie Cruise was sustained for days. We're beginning to realise that what people want to read when they are online may be very different from what they want to watch on TV.

For we now also have very rich data on the viewing of our main bulletins at six and 10 o'clock. The BBC has set up a panel of 15,000 viewers, selected to represent the population as a whole, to give their views on each evening's programmes. Amongst the questions they get asked about dramas and documentaries, viewers are also invited to select which individual news stories most stood out and which they'd like to know more about. We use this information in turn to feed the audience's curiosity about news stories. For instance the research told us, many months before politicians started to really embrace green agendas, how passionately the audience wants to understand climate change. Their interest in apparently complicated subjects like council tax and pensions comes through clearly. And, in contrast to the online audience, the current TV news bulletins audience says it is not that interested in celebrity or sensation. We know, because we also ask them what stories they are not interested in. They respond forcefully to this invitation. For example when Tom Cruise was dumped by his Hollywood studio, 80% of respondents claimed that it did not interest them. Similarly when the England squad visited the new Wembley stadium, 78% said they weren't interested.

As some of the stories the online audience loves are ones that the TV News audience says it dislikes, what should a programme editor choose to run? Some would argue that editors should just ignore confusing audience information and use professional judgement to decide what is important or interesting. And of course for the most important stories that's exactly what they do. Few audience surveys or click statistics ever indicate the Northern Ireland peace process or conflict in Sri Lanka are desired by audiences, but we need to cover those stories. But when it comes to judgments about the stories that will make up the wider range of our running orders I believe it is entirely legitimate to reflect audience interests, as well as audience responses to particular treatments and angles. But, having decided to use the information, we still have some very finely balanced decisions to make. Do we throw in a story that might be of particular appeal to a younger member of the audience, whilst running the risk of alienating more traditional viewer?