Yes, Technology Can Now Predict Hits

From the major label perspective, getting a hit typically requires two things: (1) lots of money (ie, radio play), and (2) the perfectly-receptive mass audience.

Simple game, right?  Well, it turns out the results of this alchemy can now be decoded months in advance, by monitoring early reactions across interactive platforms.

Like Shazam, the now-saturated mobile music recognition app.  Of course, songs are playing in our environment all the time, but music fans only choose to tag (ie, identify) songs that they really like.  Which, according to Shazam, means that the most proactively-chosen songs are likely to gravitate into much bigger hits – especially in the presence of serious promotion and repetitiveness.

Seriously, these guys think they have hit-prediction down to a science.  “Shazam’s Tag Charts are recognized as one of the most accurate predictive measures for the industry to see which songs and artists are resonating with people,” said Shazam Director of Music and Content, Will Mills.  “In fact, up to 85 percent of the songs that get to number one in Shazam’s charts go on to break into national top 10 charts, as well.”

So what does this radical, forward-looking list look like?  This definitely isn’t a bottoms-up, organic digital list at all.  Instead, it looks a lot like the major-driven, repetitive chart-toppers of today – and includes continued saturation of Gotye.

And with that, here’s Shazam’s predictions for the ten biggest chart-toppers of the summer.  Call BS on this now, or wait a few months for the real verdict.

(1) Gotye, “Somebody That I Used to Know”

(2) Usher, “Scream” and “Looking 4 Myself”

(3) Maroon 5 featuring Wiz Khalifa, “Payphone”

(4) Calvin Harris featuring Ne-Yo, “Let’s Go”

(5) Rick Ross featuring Usher, “Touch ‘N You” (also, “Stay Schemin'”) 

(6) Gusttavo Lima, “Balada Boa” 

(7) Rudamental featuring John Newman, “Feel the Love”

(8) DJ Fresh featuring Dizzee Rascal, “The Power” 

(9) Nicki Minaj, “Pound the Alarm” 

(10) Alabama Shakes, “Hold On”

15 Responses

  1. newm

    Gotye – Somebody That I Used To Know!?

    Has it predict Germany to lose WW2 as well?

    • paul

      Funny. Not the most shocking, but their prediction is that Gotye’s “Somebody” will continue to dominate – or is continuing to track upward.

      From Shazam:

      “Gotye, “Somebody That I Used to Know” is currently receiving 1.5 million tags each week, worldwide.”

      /paul

    • fascinating captian

      picking songs that are already hits, to be hits… LOL…

  2. Jeff Robinson

    Two great tools for artists to use now are:

    http://www.uplaya.com which utilizes the ‘Hit Song Science’ algorithm that predicted the success of Norah Jones ‘Come Away With Me’.

    Many distributors are using this service to sort through the thousands of songs they are asked to distribute. It’s funny how across the boards, the highest rated songs are consistently chosen for airplay.

    For an indie band, there is empowerment in knowing which songs have the best chance of success.

    • careful

      Last I checked uplaya’s terms and conditions (albeit it was months ago) it gives uplaya sync rights to your music, so proceed with caution on that one.

      As for shazam being ‘predictive’ I’m not sure how much of an advance this measurement tool really is. My guess is that the Big Champagne guys (now at Live Nation) might have a beef with this claim, as their tech was measuring illegal file sharing of songs as a predicitive measurement years ago – and with likely the same accuracy.

      As Paul rightly points out, these tools are more of a prescription for which songs to push by the major labels. And once you have a major label pushing a song, it can become a somewhat self-fulfilling prophecy that it will grow faster than other songs in the stable that are not being invested in as much.

      The bigger question is how it does predicting which emerging artists are going to ‘break’. My guess is that this is the 15% of top hits that it does not do well.

      And so, it really comes down to a question of ROI. If the majors use this info to invest more deeply into certain tracks, then the benefit of this info comes down to how much they get back out after recouping the investment.

      Anything else becomes correlative, not causal.

  3. Steven Corn (BFM Digital)

    To get on this list, you already have to have a lot of people tagging your songs. That means that it is already being distributed and marketed. All of these songs already have a significant level of exposure and fanbase. That’s why I don’t think that this sort of prediction is at all interesting. (I love the comment about Germans and WW2.)

    I think it’s much more interesting when a technology predicts a hit before it’s released e.g., Nora Jones and Hit Song Science.

    Now, if a technology can come up with a way to get new fans to listen and retweet an unknown indie artist’s single, that would be something amazing.

  4. Jeff Robinson

    Sorry, my gaff, forgot service number #2 in that short list-

    The second is SoundOut.com

    Paid for by broadcasters and record labels, citizen reviewers are paid to react to the latest releases. The citizens are paid by Slicethepie.com. Simply gut-level reviews of varying degrees of sophistication can give an artist a quick look at how the public at large will react to a song.

  5. @DJDangerVenture

    How did it take this long to realize this!

  6. @AJSouthgate

    Yes, Technology Can Now Predict Hits…

    …sorta

    • working musician

      With the current so called “hits” today, its pretty clear that technology better be picking them, as the masses sure have been dropping the ball on “hits” for quite some time…

      Now to figure out who is programming the technology for this…..

  7. listentogoodmusic.com

    Technology can recognize hits, but can any tool recognize good music? Money and being well received does not necessarily mean quality songs – ones we would still like to hear in the long run…

  8. aa

    sounds like self-fufilling prophecy to me. repeat a song enough times on the radio and it’ll hit top ten. instead of some dude at the label choosing which song to promote, this is marginally better, an app with a self-selecting group of listeners directing label staff where to throw their budgets.

    genius of the crowd?

    i prefer bukowski’s version.

  9. plukowski

    One way to build a ‘hit predictor’ – predict that songs that are already hits will be hits.