The quality pitfalls of user-generated content

At a conference today in Cambridge, I was struck by the thoughts of one of the audience members (which consisted of mostly ‘creative types’ - designers or artists).

I was on the panel, discussing what Virgin Radio has been doing in the user-generated-content space. A lady probably in her late 40s was concerned at what she called the ‘fad’ of user-generated-content. She described herself as a creative, and was concerned that “we were allowing” people with no creative talent to post their own work online. “Surely you have a duty to sift through”, was her argument, “to make sure that the work you promote has some artistic merit?”

She warmed to the theme. “You’re conning us all,” she said. “Do you even care what goes onto your website? It’s a totally cynical ploy to earn money out of people.”

My efforts to persuade her that the quality rises to the top wasn’t totally successful - probably because I wasn’t instantly able to show her Flickr’s automated “best of” collection, which they call ‘interestingness’.

Someone else started talking about ‘amateurs’ on the web posting content. My hackles rose; but thankfully the excellent, considerate, handsome, intelligent, erudite, and man-who-reads-this-blog, John Naughton, called time.

Michael Mullane’s blog recently featured a similar thought, describing an editor, who he calls Prendergast:

Prendergast was losing sleep over moves to embrace user-generated content. “We are pouring more and more editorial resources into sifting through pictures of kittens playing with balls of wool.”

He said that it was turning his company into the editorial equivalent of a skinny model with a disproportionately large head. “Even if we come to our senses, I’m scared that we will have irrevocably damaged our health.”

After the panel session, while talking to her along with a chap from Sky, I was struck by a thought.

You see - in many cases, quality does rise to the top - whether from popularity or from the type of algorithm that Flickr uses. (Indeed, I once wrote a similar algorithm for Media UK, which used all manner of interesting things to work out whether a particular forum thread was ‘worth reading’ or not.)

But in many cases, quality doesn’t. BBC News produces a most popular page, showing news stories which most people are emailing to their friends or reading. As of now, “Stephen Hawking In Space” is the most popular, with a story about Bob Woolmer second, and a Hugh Grant arrest in third place. Meanwhile, the US Congress has just approved a pull-out of Iraq, which the BBC bills as second-most important, but appears as the tenth-most popular story. (There’s a website somewhere which compares these two weightings side-by-side, but I can’t find it). Indeed, the second-most emailed story right now on the BBC News website is Sudan man forced to ‘marry’ goat, a story (predictably involving human/goat relations that go rather too far) which is now well over 13 months old. There’s an algorithm that needs changing.

I’ve managed to post this drivel quite successfully without anyone else checking over it for any redeeming quality - and I’m guided, predominantly, by Google Analytics (having made a conscious decision not to splatter ads all over it, otherwise I’d find myself blogging about credit cards) - so perhaps I’ve also fallen for ‘popularity is good’.

We’re drowning in content. My Google Reader is permanently full of more stuff to read. I need someone, or something, to rate that content for me - in the same way that Flickr successfully does it.

Our questioner doesn’t use the internet much, so she’ll never find this (particularly since my email address was spelt wrong in the programme, not that it would take anyone two minutes to work out what it should be). But I wonder, and whisper this, whether she might have a point after all.

UPDATE: John Naughton speaks about the panel session (MP3)