The latest radio listening figures are still rumbling round the media columns. The numbers, from the industry body, RAJAR, have generated a tremendous range of stories, even allowing for the time of year.
Radio listening is at a record high, said the Guardian, suggesting interest in the General Election and the football World Cup was why TalkSport, BBC Radio 4 and Radio 5 Live had such a good quarter. The Observer followed up by once again, pointing out that (Aha!) video had not killed the radio star. The BBC News website chose the solid, mainstream angle of Chris Evans' custody of his Radio 2 inheritance.
Good numbers for 6 Music were, of course, too tempting for anyone wanting to have a pop at BBC management. And the commercial radio lobby was (quite reasonably) keen to plug some positive news for them.
But the Guardian did a later 'think' piece pointing out that, while the number of people listening was up, the length of time spent tuned in was down. One Sunday Times columnist suggested Radio 3 figures were down as a result of too much choral music.
For anyone just coming into the journalism profession and worrying about how he or she will build a career across the converging technical platforms, this must all appear rather confusing.
The truth, surely, is that consumers are getting (as they always have and always will) more adept at fulfilling their needs and wants for news efficiently and effectively.
News in audio form will endure as long as people have to drive a car or butter toast. But, as more and more of us carry devices that will alert us to breaking news, the need to carry on listening after the current headlines have been digested goes down. We may, of course, use that time to listen to something more specialist or analytical on-demand or pre-recorded.
Clearly, the RAJAR figures have many uses and they're vital if you're working for a service with direct competitors or if you're selling advertising.
But if, as the BBC is, you're trying to meet consumers' needs and wants as effectively as possible through the full range of technical alternatives, don't you need a lot more qualitative research as well as these infinitely interpretable numbers?
