How to write a hit song
Analysis shows that charting singles are systematically different to less successful offerings. Andrew Masterson reports.
If you want to write a hit song then make sure it is happy, hits a party vibe, is not likely to relax listeners, and preferably is sung by a woman.
There. Simple. That’s the take-home message emanating from a study published in the journal Royal Society Open Science, in which a group of mathematicians analyse the success or otherwise of half a million songs released in the UK between 1985 and 2015.
Using a machine-learning algorithm known as random forests, and controlling for potentially complicating factors such as the fame of an artist, Myra Interiano and colleagues from the University of California Irvine in the US plumbed the depths of a couple of huge community-generated music databases and looked for trends developing over decades.
They also looked for successful songs and analysed their structures. Success, in this case, was, the researchers note, a “crude measure” that required only that the piece of music made it onto the UK official Top 100 Singles Chart in any year.
All of the songs in the database were classified and rated according to 12 variables, which included timbre, tonality, danceability, the gender of the singer, and whether the music could be described as conveying a mood that was relaxed, aggressive, sad, happy or likely to go down well at a party.
The overall results indicate that during the past three decades songwriters have become a rather morose and introspective bunch.
The researchers found that “‘happiness’ and ‘brightness’ experienced a downward trend, while ‘sadness’ increased in the last 30 or so years”.
This was consistent with an earlier study that found the use of “positive emotion” in lyrics had taken a dive, with recent songs more likely to focus on the self than on the concept of partnership or friends. The use of aggressive words, such as “hate” and “kill”, was also on the up.
Interiano and her colleagues also determined that the qualities of “relaxedness” and “danceability” had increased during the study period, while the percentage of male voices had dropped.
This last was particularly so when it came to analysing successful songs, which were “characterised by a larger percentage of female artists compared to all songs”.
Interiano and colleagues also determined – perhaps a little predictably – that charting songs are quite rare. Only 4% of the numbers listed on the databases enjoyed the achievement.
The researchers noted that the numbers that did crack the dizzying heights of the Top 100 singles chart were often “systematically” different from their less successful competitors. They summarised theese differences in half a dozen easy-to-remember points:
1. Successful songs are happier than average songs.
2. Successful songs have a brighter timbre than average songs.
3. Successful songs are less sad than average songs.
4. Successful songs are more party-like than average songs.
5. Successful songs are less relaxed than average songs.
6. Successful songs are more danceable than average songs.
In the first five categories, the study finds, the characteristics of charting songs run counter to the tendency towards sadness and negativity found in the remaining 96% of the tracks analysed. The researchers tentatively suggest that this could be partly because of the success of hit-packed compilation albums, or because older music consumers are introducing a level of “inertia” in the success of popular music.
The analysis also found that successful songs were likely to be less aggressive than run-of-the-mill numbers.
The fact that more successful songs were sung by women, the researchers note, was interesting in light of current debates surrounding the status of women in the music industry. These range from discussions of the sexualisation of female singers to the fact that female-led acts are under-represented at major festivals.