Songwriting: A scientific formula?

By Freya Masters
13th April, 2022

Whether I’m happy, sad, inspired, worried, or just bored… I’ll always pick up my guitar and sing.

I received my electric-acoustic guitar for my birthday when I was fifteen years old. I could go on forever telling you how resonant and beautiful my guitar sounds, because it is a ‘Martin’. Its rich sound is derived from the fact that Martins are handmade, not factory-made. As the Martin website puts it: ‘handmade guitars are special: they have a soul all their own’. 

Recently, I came across an article which concerned why scientists believe some songs are ‘great’ songs. The article talked of a ‘scientific formula’ for writing a successful song. This intrigued me. Surely there cannot be one single formula for success, given the sheer number of potential chord progressions, key changes, lyrical combinations, and melodies?

Back in 2011, the University of Bristol’s Intelligent Systems Laboratory attempted to understand the qualities of successful pop songs. Analysis of Britain’s Top 40 charts spanning the past 50 years was conducted using machine-learning algorithms. Research included quantification of features such as song length and resulted in the derivation of a ‘Hit Potential Equation’:

Score = (w1 x f1) + (w2 x f2) + (w3 x f3) + (w4 x f4), etc. 

Each letter refers to one feature of a song, such as its length. 

The equation resulted in the team predicting with 60% accuracy whether a song would enter the Top Five! 

Further attempts to elucidate the fascinating mystery of what makes a song successful included the BBC’s documentary ‘The Secret Science of Pop’, which involved researchers here at Imperial. More recently, machine learning was employed to analyse 80 000 chords belonging to American billboard hits. The impact of feeling uncertainty or surprise on the brain’s pleasure when certain chords were heard was measured. It was found that the surprising chord progressions were more effective at stimulating pleasure in participants, with patterns of brain activity observed during functional MRI (magnetic resonance imaging) indicating this response. 

If you want to read more about these studies, links to relevant papers are located at the bottom of this page. Furthermore, there is a link to Lia’s article which details how exactly we experience pleasure in the brain from music…

Contemplating all this, I wondered whether I use a formula in my own songs. It occurred to me that I do, in a way. I always write the chords first (mostly minor chords), and after I’ve established the chord sequence, I start to hum a melody. From this humming, I write my lyrics. I’ve always done it this way. Not a complex equation really! However, I decided to shake up my ‘formula’ recently and wrote the lyrics first before the chords. Madness! 

In the end, perhaps there is a beauty in scientists not deriving a formula for what makes a good song. There is pleasure in not knowing, merely listening. After all, there are limits to what science can achieve. I perceive songs as positioned at the wonderful intersect between science and the arts. The mathematical basis of a song is entwined with the wonder of writing its lyrics and cultivating the feelings it evokes within. There is pleasure in writing songs as and how chords emerge from your fingertips, and words from your mind – without focusing too much on why. 

---

2011 study: http://www.tijldebie.net/system/files/MML2011-final.pdf

2019 study: https://www.sciencedirect.com/science/article/pii/S0960982219312588

Why tiny vibrations in the air have so much meaning in our lives: http://isciencemag.co.uk/features/music-why-tiny-vibrations-in-the-air-have-so-much-meaning-in-our-lives/

---

Freya Masters is the Online Features Editor for I,Science and is enrolled on the MSc in Science Communication at Imperial College London. She loves writing in all forms, with a passion for science writing in particular.