How Big Data Influences the Pop Song Writing Process

And the beat goes on...and on. JL

David Plantz comments in Quora:

What's the average song length? Intro length? Structure? Start with chorus? Is there prechorus? Most common instrument? Are male vocals or female vocals doing better? How long are echos and reverbs? Digital drums or analog? Acoustic versus electric guitar? Most popular key? Major or minor?...

Advertisement

Data is certainly popular enough to sustain businesses like Hit Songs Deconstructed, which writers, producers, record label, and radio executives subscribe to. Just like in other parts of the entertainment industry, following music trends is important to economic success, not only for executives but also for artists.

Let me clarify what I mean by music trends. It is more than the genre discussions most of us find ourselves in from time to time: Is dubstep still relevant? Did hip-hop jump the shark? Is disco back? Is country too pop these days? Will their ever be a next Nirvana?

Song writers and producers want to know even more: What's the average song length? Intro length? Structure? Start with chorus? Is there prechorus? Most common instrument? Most common secondary genre? Are male vocals or female vocals doing better? How long are echos and reverbs? What frequency range is most emphasized? Digital drums or analog? Acoustic versus electric guitar? Noticeable effects? Average BPM? Swing in the tempo? Most popular key? Major or minor? Popular lyrical directions? Snare hits on beat two or beat three? Which songwriters are crafting most hits? What genre is out of top 20? What genre is rising in past six months? And there's so much more ...

I've had access to this data before and wrote a pop song or two based off it. Quite honestly, the songs I write with data objectives haven't been as good as the ones I write for fun. One would think all this data would make it easy. It's not. You have to be able to balance trends while still committing to something. It's more important to know when to use the data and when to ignore it. And of course, the song still has to be good, attention-getting, familiar yet distinctive, and memorable. On top of that, it has appeal to the artist on his or her own merits and fit  his or her image and audience.

If there's anything people should know about pop song–writing is that these days the biggest hits are collaborative team efforts. The days of Brian Wilson crafting a top 10 hit are rare, at least for now. You'll have one person craft the beat, another craft the chorus and foundation, and another craft the verses. Then the lead producer will work with the team to bring it all together into something that meets the artist's and the label's visions. (This is when things like guitar solos are nixed for length, tempos are adjusted, secondary genres are determined or change, etc.) The mixing engineer and the mastering engineers bring their expertise as to how to make something sound more modern or more retro based on the technical trends (frequency analysis, song dynamics, depth, etc).

So yes, data can and does play a big role. But there's still an art to pop. Otherwise, we could just have machines craft the hits. Maybe down the road, though.