The integration of big data into something as simple as music listening is a far more invasive process than we would originally consider. Now more than ever, we are seldom without the effects of an analyzed dataset telling us what to do.
It’s hard to imagine a time when big data hasn’t influenced the music industry. In nearly every avenue consumer data and behavior have been exploited to create what can only be considered a whole new era of music listening. The cultural zeitgeist has always been on the minds of artists and studio executives alike, but what happens when consumer data is the only metric taken into account for creating user experiences on major platforms?
There are a variety of ways in which big data finds its way into your habits and consumer behaviors. Every retail company has a “predictive analysis” department devoted to understanding its market by analyzing its personal patterns and shopping conventions. What most people don’t understand is that music is a business, not just an art form. Analytics and data are just as important to the music industry as to any other retailer on the market. Streaming services want to keep your subscriptions and artists want you to listen to their music. They’ll go to great lengths just to achieve this.
Nowadays music listening has been completely streamlined. Patrons only have a select amount of avenues to listen to music but a greater amount of choices than ever before. So, what’s the solution to this pressing dilemma? How do you get people to want to listen? The answer comes in the form of big data. Big data in the music industry has highly affected the way in which listeners discover, consume, and engage with music. Subsequently, changes in music marketing and even how artists achieve success are all consequences of its latent usage.
Music Discovery and Playlist Curation
Technology has been one of the biggest factors that have influenced the industry at large. However, it’s not just technology that creates these complex and robust algorithms designed to keep the listener engaged. Nick Seaver has a fantastic book on big data and the music industry called “Computing Taste” which goes in-depth on how algorithms are determined and created for people by people. It’s a process that touches areas of study like psychology and even neurology in some instances. One of the terms the book speaks on is called the “hang-around-factor.” It’s a blanket term that essentially describes what factor is going to keep the listener engaged and listening to any recommended song.
If you’ve listened to a song before that was recommended to you and within the first few seconds you instantly skipped it, people working the algorithm stop at nothing to understand why you made that choice. They want to create a butterfly effect of sorts and see where every road meets. Every single change needs to be scrutinized for its relative impact, positive or negative.
The main idea of such an algorithm is to inevitably get you hooked on whatever a given station is playing. The information age that we find ourselves in today will really only get you to the tip of the iceberg. However, the market is driven by socio-economical factors that inherently benefit from a consumer’s willingness to be eclectic. Algorithms have to essentially trap their consumers, they have to convince you to want to stay and listen. The most important thing to any algorithm is whether it can captivate the listener or not.
Seaver goes into more in detail on how recommenders are actually able to calculate what someone wants to listen to by comparing “predicted values” and “actual values” to create “ root mean squared error:” While we won’t be getting into that it’s still a concept worth noting because of it’s an inherent connection to the metrics that algorithms use. “RMSE” can only be understood in isolation and can easily fall to time, they are simply just sets of data that only exist in a flashbulb memory. Seaver mentions an alternative to the RMSE known as “captivation metrics” which have the ability to measure over time and outside of the usual confines RSME has.
“Captivation Metrics” are invisible a lot of the time but also tangible actions taken by any given consumer, in a way they’re a lot like how we think about data colloquially. This is data generated from retention rates, to repeated listening all the way to tracking your basic interactions with a given medium. They measure the level of engagement and attention that a piece of content and they’re so important because they make a bigger splash due to their retention. It’s essentially a feedback loop of sorts using suggestions crafted to encourage engagement, swiftly boosting your visibility in the logs. It’s a rather simple ideology, the more time you spend creating data for someone to study the more important of a consumer you are.
The only downside to such a business model is the lack of individuality. While you may think that your curation is completely unique to your person there’s an intrinsic and most definitely intentional consequence to relying on them. Ultimately, it leads to a narrowing of musical tastes and a lack of diverse music consumption. You are unlikely to be exposed to new or niche genres and encouraged to homogenize. Moreover, algorithms tend to prioritize artists that generate the most streams and buzz. This is how big data has affected music marketing, artists now have to conform to established conventions and play to trends in order to gain visibility.
Music Marketing and AI Generation
To be able to market your music effectively, you have to paint a picture of what your audience is like. Whether you can mold them is another story. The beginning is always the hardest part because of the lack of available data present. However, the steps following are all there to learn about who is listening. To create a profile and use it for you, or against you. So what are listeners actually like? It’s a question once again posed and answered by Seaver.
Music is marketed not necessarily to a person but rather an idea, an idea based on a realm of factors predetermined by those who created the image in the first place. The purpose is to bridge the gap between the listener and the process of captivating. This is the part where you’re meant to become a part of the process. In nearly all cases, a listener will rarely coincide perfectly with the image marketed to you. However, there is meant to be an inherent realm of relatability that makes it nonetheless easy to identify with.
Nowadays recommendations and viewers of the listener are what Seaver describes as “post demographic.” Something that is meant to move away from traditional demographic categories, such as age, sex, race, and gender, and toward a more nuanced understanding of consumer behavior. This trend in algorithms has been made possible by the wave of big data. Music marketers desire a more curated and unique strategy to appeal to their audiences, however, despite this, there is technically no way to divorce from traditional demographic matrixes. Despite their best efforts, music marketers understand this as well which is why marketing has always included some appeal to tried and true demographic factors.
This has also created a shift in how artists achieve success, with record labels relying more heavily on data to identify emerging trends and new talent. One of the most ironic trends of recent years has been AI-generated artwork. American rapper Lil Yachty recently released his fifth studio album Let’s Start Here which features an AI-generated album cover. The cover itself isn’t too outwardly appealing, it depicts a businessy crowd of people with crooked and warped smiles. It’s supposed to emphasize the uncanny valley effect their faces generate and it received a lot of media coverage for being one of the first popular contemporary albums to use Ai-generated artwork.
There’s been a lot of questions on whether this is a genuine form of artistic expression but the most intriguing part about its usage has been its indifferent integration into such a popular medium. It lightly signals our growing desensitization of big data in our everyday consumer practices. It’s fun to poke and laugh about eerie and strange internet artwork but we don’t realize our growing cultural numbing to these concepts. However, that’s the entire point.
Big data wants to be invisible, it wants to be this character that falls just outside the line of comprehension. It’s often difficult to grasp due to its sheer size and complexity, the vast amounts of information, and cutting-edge tools. It’s always at work, shaping our experiences and influencing our decisions. By keeping itself in the shadows, it’s able to exert a widely felt impact on a wide range of industries, including the music industry, without being thoroughly scrutinized.
Lorraine Daston wrote in her essay “Objectivity and the Escape from Perspective” that objectivity is a concept that is hard to disconnect from “point of view” and it would seem big data has made that harder.
A user determined entirely of ones and zeros is just a black mirror reflecting back their own preferences and habits. It cannot stand in or substitute for the human making the decisions. What big data is unable to replicate or communicate are the deep characterizations that lead human decision-making in the first place.
Humanity must ask themselves if they’re ok with the way something as pure as music is so heavily integrated with the capitalist nature of big data. We aren’t just selling a product anymore, in a lot of ways we have become to the product. Part of the industrial chain continues to move at an indefatigable pace.
Daston, Lorraine. “Objectivity and the Escape from Perspective.” Social Studies of Science, vol.
22, no. 4, 1992, pp. 597–618. JSTOR, http://www.jstor.org/stable/285456. Accessed 20
Duhigg, Charles. “How Companies Learn Your Secrets.” The New York Times, The New York
Times, 16 Feb. 2012,
Seaver, Nick. Computing Taste: Algorithms and the Makers of Music Recommendation.
University Press of Chicago, 2022.