KonzeptioNerd

Dusting the Crystal Ball

Kate Green / Technology Review reports on an HP Labs initiative to calculate the future success of online content (Die Popularitätsformel, Dec 30, 2008). The researchers at Palo Alto look at the first day metrics of online text or video: Digg and YouTube primarily., apply some mathematical models, and come up with a popularity curve for a given content. This curve shall then help determine the value of adjacent ad-spaces. Two questions.

First, I wonder if this can work for non-English language content. German content rarely gets critical mass on Digg, delicious or Technorati. Can we factor this effect in with a stable parameter? Which would it be? Ratio of German-speakint to English-speaking internt users worldwide? What about alternative metrics? There’s Mister Wong for delicious, but not much else. And what would we get? A popularity curve across German speaking communities. Is the HP curve, then, global? Or US centric?

Green quite rightly draws in Claudia Perlich of IBM Watson Research Center to caution that, when the success of advertising is at stake, it’s not enough to place advertising next to popular content, one also needs to know a bit about the visitor (e.g. her clickpath leading up to this piece of content) in order to target properly. But if you start targeting, shouldn’t well targeted niche-content be more valuable than globally popular stuff?

Second, the article suggests (but doesn’t state quite clearly), that the popularity curve is platform-specific: for Digg, it’s a couple of hours, for YouTube it’s 20 days, for online shops its particular and as yet undetermined. That looks like a pretty linear relation. Too bad Digg doesn’t supply popularity graphs for individual stories. Or YouTube for videos. So the question is: how did the researchers arrive at their numbers?

The other questions is, wouldn’t I rather either content-target my adverts (if I can) or simply place them next to whatever is (most) popular at a given point in time (*if* this is platform-specific)? 

And, at any rate, as an advertiser, I’d be more interested in figuring out WHAT makes content popular (even within the context of individual platforms) and then build my branded content accordingly.

Leave a Reply