EMERGE's big brother blog, Influx Insights wrote this article last week praising Malcolm Gladwell's recent article in the New Yorker.

The piece was a lengthy and indepth look at a couple of new algorithmic analytic tools used to predict consumer response to potential film scripts and pop music. While the article does discuss the reluctance of both the film and music industries to wholeheartedly embrace such "formulas" for success, it only devotes a paragraph or two amongst several full pages of text to notions of "of something called that magical thing- talent, the unexpected."

However, EMERGE remains skeptical against Gladwell's mountain of research and quantitative evidence. But rather than being content with our own doubts, we put the question to a group of roughly 500 music industry experts and insiders who all belong to a Yahoo! group called The Mishpucha to see what their thoughts were. While one of Gladwell's subjects, Mike McCready, was in fact one of the mishpuchim for a while, he was drubbed off the list by continual derision from hundreds of highly articulate music critics who disdained his purported succes at "figur[ing] out how the brain works regarding musical taste." Yes, EMERGE knew it'd get some juicy rebuttals by asking these folks! In fairness though, we tried to get McCready involved. Unfortunately, repeated emails went without response, though some of his detractors laid out some very good points quite quickly...

One respondent, who works as head of mobility content for a major mobile handset manufacturer stated simply: "A "hit song" is just more of the same old bullshit you always listen to??? I think this gold rush is more than a little misguided." While her sentiment is widely shared amongst the members of the "Mish", music critic Douglas Wolk made a more detailed critique, using his encyclopedic musical knowledge to debunk McCready's central claim in Gladwell's piece. Rather than distilling the essence of what he said, EMERGE would rather share Doug's response in its entirety:

"You know, I looked into that stuff a few months ago for an article
that never happened for various reasons, and I'm surprised to see
Malcolm Gladwell, of all people, buying into the hype
uncritically--because neither his article nor any other press I've
seen about Hit Song Science or Platinum Blue gave any suggestion of
what kind of "predictive" powers they've got, or what sort of false
positives or false negatives they generate. I mean, it's easy to say
ex post facto that your software gave "Crazy" a "Hit Grade" of 755;
maybe "SexyBack" got a failing "hit grade," maybe some heavily
promoted major-label thing that stiffed anyway got a "hit grade" of
900, but we don't get to find out about that.

Also, Gladwell says that nine of the fourteen songs on Norah Jones's
"Come Away With Me" had "clear hit potential" (according to "a firm
in Barcelona" which remains unnamed in the piece--it would be Hit
Song Science, btw); curious, then, that the only actual hit on it was
"Don't Know Why," which only made it to #30. That record's not about
being packed with hits, it's about an overall sound and style and
persona. And anybody who thinks that hits--big, popular hits--are not
_significantly_ about image and persona and stuff that has nothing to
do with .AIFF files is dreaming."

A pretty pithy critique and one that probably won't have McCready racing to resubscribe to the Mishpucha group any time soon. While it doesn't address Epagogix, who's script analysis software, the sentiment of the Mishpucha still seems relevant, especially since Gladwell's subjects have literally written the cast, director and special effects right out of the analysis, leaving just the screenwriter's vision of the film to be analyzed. Not exactly the sort of rigorous look-at-every-possible-variable analysis that would satisfy the skeptics among us.

However, one component of these predictive technologies which Gladwell didn't mention nevertheless got some praise from our informal panel of popcultural know-it-alls. Apparently, Hit Song Science have recently changed tack, working to integrate their software analysis into recommendation engines to augment the traditional model which bases suggestions on buying history and the histories of other users. Perhaps this is the sort of thing NetFlix should be doing rather than providing anonymous copies of it's entire user database to nerds worldwide to compete for a hefty purse for improving their inhouse recommendation software.

It appears you don't have Flash installed.