The age of average (encore)

Preface

In March 2023, I published The Age of Average. The article argued that the world in which we live, including its art and architecture, film and fashion, increasingly looks alike.

Since publishing the piece, it’s been read over half a million times, presented at conferences and discussed on podcasts.

But despite its success, there’s a chapter that has never seen the light of day. This chapter delved deep into how our musical culture is homogenising in exactly the same way as our visual culture.

Now, I’ve finally found the time to turn this “lost” chapter into an essay of its own.

Here goes.

Introduction

Media has always shaped music.

In the 1950s, the limited capacity of 45rpm vinyls created the conditions for the 3-4 minute pop song. LPs, which held 22 minutes per side, ushered in an era of roughly 12 track albums. And radio’s popularity led to long, instrumental introductions that disk jockeys could talk over.

Clearly, how music is distributed, affects what music is distributed.

In 2005, streaming services made up a negligible amount of recorded music revenues. Today it accounts for over 80%. We are now 15 years into the streaming era, where anyone can listen to any track at any time. And whilst it was predictable that music itself would evolve, how it would evolve was anything but.

This article argues that the changes in technology have led to songs getting shorter, music getting less melodically diverse and lyrics getting more repetitive.

Or to put it another way, just as our visual culture has become more homogeneous, so too has the music that accompanies it.

Let’s run through these arguments one by one.

Music is getting shorter

In October 2022, Lil Yachty uploaded his track “Poland” to SoundCloud. The song, which consists of two keening hooks and some slack rhymes was a hit. It reached number 40 on the Billboard Hot 100. The whole track, start to finish, lasted only 83 seconds.

This brevity might seem unusual. But it is, in fact, becoming a new normal. Some of the biggest hits of the last few years, such as Lil Nas X's Old Town Road and Nathan Sykes' Wellerman, have clocked in at under two minutes. And in 2016 Piko-Taro released “PPAP” which, at just 45 seconds, became the shortest Hot 100 entry ever.

Research shows that these tracks are examples of a much broader trend. According to Elias Leight, writing for Billboard:

“The average length of popular songs has been shrinking steadily for years. A 2018 study by San Francisco-based engineer Michael Tauberg concluded that songs on the Billboard Hot 100 shed around 40 seconds since 2000, falling from 4:10-ish to roughly 3:30. The average length of the top 50 tracks on Billboard‘s year-end Hot 100 in 2021 was even less, a mere 3:07.”

Leight goes on to state that around 4% of top 10 hits were sub-three-minutes in 2016. By 2022, that figure was 38%.

And this trend isn’t confined to the US. We’re seeing it in the UK as well:

“A study by Manchester-based digital label Ostereo found that the average duration of number one songs in the UK in 1998 was 4 minutes 16 seconds. In 2019, that average was down to 3 minutes and 3 seconds with 2018 being the year with the biggest drop.”

Whilst lazy reporters put this trend down to the TikTok generation’s faltering attention spans, the truth is far more prosaic. Shorter tracks make more money.

Streaming platforms pay artists each time a track gets listened to. And a “listen” is classified as 30 seconds or more of playback. To maximise their pay, savvy artists are releasing albums featuring a high number of short tracks. In purely commercial terms, an album with 20 two minute tracks will generate double the revenue, per play, than an album with 10 four minute tracks.

Here’s an excerpt for The Verge’s Switched On Pop podcast:

“One of the main things that has changed is how people are getting paid, and it’s affecting how songs are being written. (…) Instead of getting paid by physical sales, you’re getting paid in a stream, which only counts if someone listens to 30 seconds of a song. It actually makes sense if you can have more songs streamed at a time, which means that you want to pack your album full of much shorter songs. So if you have an album like Drake’s Scorpion, which is a really long double album coming in at almost 90 minutes, he’s got a ton of really short songs on there, because he gets paid for every song you listen to, whether or not you listen to the whole album.”

In the age of streaming, songs are getting shorter.

But this isn’t the only way pop music is changing. 

Music is getting less melodically diverse 

As songs are getting shorter, their structure is changing to suit. The most apparent change is the disappearance of the intro.

Research by musicologist Hubert Léveillé Gauvin found that the average length of introductions has decreased 80% over the last 30 years:

“Léveillé Gauvin looked at 303 top-10 singles released between 1986 and 2015. He found that the anatomy of pop music had changed significantly, especially in terms of the length of the instrumental introduction. In the mid-1980s, the intro ran 20 to 25 seconds. By 2015, it had shrunk down to a mere five seconds.”

Léveillé Gauvin describes the effect as “front-loading”. To achieve the critical 30-seconds of playback, lengthy intros have fallen out of fashion, and choruses increasingly break in the first few bars. 

This “front-loading” effect is also affecting music’s ‘final third’ as well. 

An analysis of the Billboard Top 100 by Chris Dalla Riva found that the use of key changes has fallen from around 30% of tracks in the 1990s to low single digits today:

“The act of shifting a song’s key up either a half step or a whole step (i.e. one or two notes on the keyboard) near the end of the song, was the most popular key change for decades. In fact, 52 percent of key changes found in number one hits between 1958 and 1990 employ this change. (…) What’s odd is that after 1990, key changes are employed much less frequently, if at all, in number one hits.” 

And it isn’t just introductions and key changes. Hannah Davies, at The Guardian, has argued that bridges are dying a death as well:

“If you have listened to pop music at all in the past few years, you may have noticed that something is missing. The bridge – that part of the song where verse and chorus give way to an alternate section that ramps up the tension (or the fun) – is seemingly on the wane. Where a big, barnstorming section (…) may have once been de rigueur, these days you often just get another verse or a moody final chorus.”

The net result of intros, key changes and bridges all falling out of favour is a measurable reduction in music’s melodic diversity.

In 2005 Tristan Jehan, a PhD student at MIT, published his dissertation in which he calculated the 8 key dimensions of music. 

Shortly after completing his dissertation, Jehan co-founded EchoNest where he graded songs on each of his 8 dimensions. Spotify quickly saw the value in EchoNest’s work and used the dataset to fuel its recommendation engine. If a listener showed signs of liking a track then Spotify would recommend other songs that had a similar sound signature.

Essayists at The Pudding then used this data to calculate the average difference between a charting song and all others in the Hot 100 at the time.

The study shows a gradual but consistent decline in musical diversity since the early 70s:

“The result is a trend toward similarity, with smaller distances [in EchoNest scores] among songs. To date, songs that charted between 2012 and 2016 were the most similar, [of any era measured].”

Between 1958 and 1961 the average distance between songs was 73%. Between 2012 and 2016, the score had fallen to 56%. Melodic diversity had fallen by almost a third. 

In short intros, key changes and bridges are used less frequently and, as a result, songs are getting less melodically diverse.

But this isn’t the only way pop music is becoming more alike. 

Music is getting more repetitive

In 2017, American Rapper Lil Pump released his self-titled album’s fifth single, “Gucci Gang”. The track peaked at number three on the Billboard Hot 100 despite the song’s length being just 1 minute and 22 seconds. Whilst its brevity is no longer unusual, there was something else that made the track stand out.

Throughout the course of the song the title is repeated 52 times. That’s once every 1.57 seconds. In fact, a quarter of all the words in the rap are either “Gucci” or “gang”.

Is this repetitiveness unique to Lil Pump? Or is the track indicative of a wider musical trend?

The computer scientist Colin Morris set out to answer these questions using a database of 15,000 songs’ lyrics that have charted on the Billboard Hot 100 between 1958 and 2017.

To distil a song’s lyrics down to a ‘repetitiveness score’ Morris used the Lempel-Ziv algorithm (the lossless compression algorithm that powers the .zip archive format). Lempel-Ziv works by pairing down repeated sequences of characters. In layman’s terms, the algorithm would discard 51 instances of Lil Pump’s “Gucci Gang”, and only save one. The more the algorithm can compress a song's lyrics, the more repetitious the lyrics must be.

The most repetitious song of all time, according to Morris’ ingenious technique, is ‘Around the World” by Daft Punk, released in 1997. The track can be compressed by 98%, going from 2,610 characters to just 61. In fact, its compressed lyrics could fit inside a tweet. Twice.

The second most repetitive song in the dataset is The Rockerfella Shank, released by Fatboy Slim in 1998. The track’s lyrics can be reduced in size by a massive 95%.

Beyond these two tracks from the late 90s, if we look at the league table of the most reducible, repetitive Billboard Top 100 songs of all time, a further four of the top 10 were released after 2010. 

According to Morris, in 1960 the average song was 46% compressible. By 1980, that figure had reached 49%. By 2000 it was 51%. And by 2015 it reached 54%. Amazingly, tracks that reached the top 10 in that year had an average compression score just shy of 60%.

Morris’ data shows that songs are clearly becoming more and more repetitive over time:

“2014 is the most repetitive year on record. An average song from this year compresses 22% more efficiently than one from 1960.”

Another way of looking at Morris’ dataset, is to assess it by artist rather than by song.

Across the 60 years of music analysed, Rihanna had the highest average repetition score (65%). Jason Derulo came in second (62%) followed by Maroon 5, Lady Gaga and The Weeknd (61%). Beyoncé, Britney Spears and The Backstreet Boys came in next (60%) followed by Avril Lavigne, Nickelback and One Direction (59%). By now, it should be obvious, but every single one of these acts found fame after the turn of the Millennium.

Evidently, songs are not only getting shorter and less melodically diverse over time. They’re also becoming more lyrically repetitive.

So where does this leave us?

Conclusion

So there you have it. Chart topping tracks are becoming shorter, less melodically diverse and more lyrically repetitive.

But perhaps we shouldn’t be surprised.

Research has found that a smaller number of superstars make up a greater share of ticket sales, that each year the top 100 chart is made up of fewer artists, and that fewer songwriters are responsible for writing the tracks that put them there.

In other words, fewer producers are writing songs for fewer artists to perform. No wonder everything sounds the same.

But this homogenisation presents an opportunity for the creatively courageous.

I believe that throughout history, the greatest musical leaps have been made by artists who chose to turn away from the tried and tested.

Bowie constantly evolved his identity; from “Major Tom”, to “Ziggy Stardust”, to “Aladdin Sane”. Bob Dylan shocked his folk fans by releasing an electric album. Queen’s “Bohemian Rhapsody” shunned the three minute single for a six minute suite combining a ballad, an opera and a rock song. And the Beatles moved swiftly from rock to psychedelia and from Eastern influences to the avant garde.

Music, unlike marketing, is not about repetition. It’s about revolution. It’s not about consistency, but conception. Not iteration, but invention.

The shift to streaming has caused the industry’s commercial interests to outweigh its creative ones. To regain its role as a cultural force music must break its boundaries, not abide by them.

So let’s take the record off repeat. Let’s stop the needle from skipping. Let’s stop waltzing wistfully through our age of average.

Let’s make music as disruptive as the technology that distributes it.

Notes

  • A big thank you to Toby Ososki and Matt O’Dowd for their guidance and support on this essay.

  • Research by Joan Serrà, of the Spanish National Research Council in Barcelona, found that songs have become less varied in terms of timbre, pitch and loudness over time. Thank you to Mike Ryan for sharing this insight with me.

Previous
Previous

Thinking rationally about emotion

Next
Next

The age of average