Opinion: music has become more of a science, relying less on human skills of curation and intuition and more on algorithms and consumer data

We listen to music for a number of reasons. We listen to alleviate boredom and pass the time, to improve productivity at a given task such as work or exercise and in socialisation and ritual. What unites these disparate uses is that the end result, intentional or otherwise, tends to involve some movement in mood and emotion.

It can be also seen with something as monotonous as muzak which is made with the intention of creating an audio brand or audio environment in order to sell or incite an emotion. The term "muzak’, associated with bland elevator music, comes from an actual company formed in 1934 called Muzak.

Muzak specialised in curating background music for businesses and retailers and was responsible for the use of terms of the highest bolloxology (in this article!) such as "audio brand" and "audio environment". Although often mocked, Muzak was at forefront of developing a lucrative industry that continues to grow today with the entry of more sophisticated big data start-ups. The global market is now estimated to be worth anywhere from $2-4 billion.

From 2fm's Chris & Ciara Show, an interview with Jack Strattan who has released an album of silence on Spotify.

The growth of companies such as Muzak from the 1930s on was another example of an increasing trend for businesses to outsource various functions and tasks. When we think of outsourcing, we think of businesses transferring control of various functions to specialised individuals and/or organisations. This has been a continuing trend since the industrial revolution and isn’t likely to change any time soon as we enter an age where tasks are increasingly outsourced to automation. 

Outsourcing is based on simple principles of economics and practicality. Consequently, because we associate this type of compartmentalisation with such mechanical terms, we often overlook how businesses also outsource emotional functions (e.g. customer services) and emotional labour (employees).

The same thing is happening with music. Muzak and contemporary successors such as Soundtrack your Brand and Mood Media are contractors that businesses outsource to fulfil various mood and affect related objectives. The outsourcing of music has become privy to similar processes of automation and machine learning as other outsourcing activities. Background music has become even more of a science, relying less on human skills of curation and intuition and more on algorithms and consumer data. 

"Of course, we have looked to music for solace in this way and to achieve specific moods and affects long before Spotify"

This can have a big impact on sales. For example, researchers from the Swedish Retail Institute  and Soundtrack your Brand were able to demonstrate a nine percent increase in sales in McDonald's when music that matched the brand (music that was welcoming and modern, apparently) was playing, compared with random music. Hundreds of thousands of tracks are analysed to find the exact song that match the parameters set by a brand such as brand values, demographics of consumers and desired ambience.

This careful curation of music for corporate mood-related goals by human or machine is also something that is prevalent in our individual consumption of music. Spotify’s value as a company is essentially built on its discovery and recommendation functions, fuelled by the large amount of data it collects from its consumers every day at various touch points from what artists you listen to most, how long you listen and when etc.

Music is no longer chosen primarily by artist, but by mood and affect

This data is then linked to wider information concerning how our individual consumption of particular artists and genres is understood. This is done through music data platforms such as The Echo Nest (now owned by Spotify) which synthesises billions of data touch points (e.g. the tempo, lyrical content and even the ‘valence’ or happiness of a particular song) to automatically classify audio content and predict likelihood of a particular listener responding to a particular song.

As a result, the playlist has become king. Music is no longer chosen primarily by artist, but by mood and affect. Spotify is littered with playlists that are designed to attain the appropriate mood for a specific task or situation from gym workouts, travel, Monday motivation and chill out mixes to "Songs to Sing in the Shower" (over three million subscribers) and even "Coping With Loss", a playlist designed to help you deal with the death of a loved one featuring such acts as Eric Clapton, Heathers, U2 and Snow Patrol.

Of course, we have looked to music for solace in this way and to achieve specific moods and affects long before Spotify. However, this was much more spontaneous and involved a greater level of labour and sense of involvement on the part of the listener. Importantly, it was a mostly human-controlled experience. The specificity of the playlists available to the contemporary music consumer and the continuing improvements in the recommendation functions mean that we can use music to experience desired feelings in a very direct and an almost mechanical way. It is almost as if we are outsourcing our emotions to the algorithm in much the same spirit as we have discussed with organisational outsourcing and corporate muzak.

What does this mean for our relationship with music, as streaming services continue to grow in market share and machine learning continues to advance at an exponential rate? On the one hand, there is a clear unease at the compartmentalisation and functionalization of music listening. We have long derided muzak because of how it essentially turns music into wallpaper for corporate means and by proxy deadens the tension that is so crucial to the emotional experiences we look for in art. There is also an argument that this can lead to less experimentation or more predictable music as artists look to adhere to the trends of what the big data is telling them to capture an audience on increasingly competitive streaming platforms.

It is also important to ask the question: what is actually wrong with emotional outsourcing?

On the other hand, one could contend that we can use the improving accuracy and creative introspection of the big data to create this unpredictability and generate a thrilling and meaningful experience of music that addresses some of the issues with consuming music in this way. There is a strong argument that this is happening already as individuals learn to navigate the use of such new technologies in different ways.

Furthermore, it is also important to ask the question: what is actually wrong with emotional outsourcing? If it works efficiently for organisations, perhaps it can help us manage mood in our everyday lives better and become more efficient and possibly better people. The problem with this argument, however, is that combining the words "music" and "efficient" doesn’t exactly set the pulses racing.


The views expressed here are those of the author and do not represent or reflect the views of RTÉ