Going Viral: What One Week in March Can Tell Us About Popularity on Spotify
By Cope Fitzharris, Maxine Malvar, Luke Slomba
One of Spotify’s stated goals is to democratize the music industry by allowing more and more artists and songs to have a place on its platform. An astonishing 60,000 songs are added to the platform every day, offering artists the chance to gain exposure and build a fan base. The number of songs and artists on the platform is increasing at a swift rate, with Daniel Elk, Spotify’s CEO, stating that as many as 50 million artists (plus podcasters) could be on the platform by 2025.
However, numerous critiques have been levelled at Spotify claiming that it has only exacerbated inequalities in the music industry. Spotify and the major record labels have benefited the most from this transformation of the music industry. First, Spotify takes home a percentage of its monthly distribution fee, and then distributes the rest to the labels . The labels, who receive around 52% of the revenue for each stream, pay artists a royalty that can fluctuate from 15 to 50 percent. Furthermore, 78% of all music that is streamed on Spotify is controlled by three record labels plus Merlin. While bigger artists are insulated from these inequalities, everyone else in the music industry is not so well protected.
Only a small number of artists earn more than $100,000 from their streams on the platform. There are approximately 800 artists who earn more than a million dollars from the platform on a yearly basis, and a further 7,500 artists who earn more than $100,000 per year. With around 8 million creators on the platform, these figures demonstrate how difficult it is for artists to earn a sizable income and break into the upper echelons of the music industry. Given this unequal situation for the vast majority of artists, our interest lies in figuring out whether or not a small number of artists are represented when it comes to which songs are being streamed the most in a certain country. The extent to which the top songs are streamed relative to songs below those top songs are also pertinent to our investigation.
The What, Where, When, and Why
In order to do this, our group analyzed data from Spotify Charts, originally collated by Spotify, that lists the top 200 most streamed songs from a single week. The dataset we are analyzing specifically is from the week of 3/4/2021–3/11/2021 and is a sample set of nine countries chosen at random — Vietnam, Turkey, USA, Mexico, Morocco, Poland, Sweden, Japan, and Egypt. A total of 1,800 entries were included in this sample, listing the song name, artist, number of streams, position within a country’s Top 200 chart, and respective countries for each placement. It is important to distinguish between the number of entries and the number of songs total; though there are 1,800 entries, there are naturally some songs that appear in multiple countries’ Top 200 streamed songs for that week. For example, in our initial analysis of the dataset, we generated descriptive statistics and all received a different song as the most frequently listed Track Name since three songs (“driver’s license” by Olivia Rodrigo, “Blinding Lights” by The Weeknd, and “Save Your Tears” by the Weeknd) appear on all nine countries’ lists.
In analyzing Top 200 streaming charts, it is important to identify significant factors that lead to placement within this top 200, like the Spotify playlist called New Music Friday, which is curated entirely by Spotify editors with no direct influence from the platform’s recommendation algorithm. A study from UC-Davis found that of distinct songs recommended on this playlist in 26 countries, almost a fifth went on to appear on a country’s Top 200 chart. Rankings on New Music Friday also affect the Top 200, with 85% of the songs ranked #1 on New Music Friday breaking into the Top 200 chart. Recognizing one of the arbitrarily created playlists and its effect on our dataset is integral to our full understanding of the dataset, and provides additional insight on the imbalance of the music industry.
An important caveat to analyzing this dataset is that this data comes from one platform, from a single week, and from only 9 of the 93 total countries that currently use Spotify. Additionally, there are 33 different music streaming services worldwide — the majority, including Spotify, are not actually available globally — and the popularity of each service is indubitably variable across different countries. Given both of the above, any findings from this project are not meant to be representative of the entire music industry, but a preliminary analysis of certain countries’ music streaming data.
Unique Artists
Our objective in analyzing this dataset was to evaluate the accessibility of an artist to Spotify’s top charts. An initial method for interpreting the data is by analyzing the number of unique artists represented in the Top 200 chart for each country. We can do this by creating a filter within the dataset that isolates each country’s chart and then removes repeated entries under the Artist column. For example, the boy band BTS had twelve songs in Japan’s Top 200 chart for this week, so eleven of these entries were removed by this filter. Through this method, we found the data plotted below:
This graph reveals the disproportionate representation of artists in each of the nine Top 200 charts we analyzed. The number of unique artists steadily declines with fairly uniform differences across countries. The average number of unique artists placed on a country’s Top 200 chart sits at about 109, which indicates that an average of 91 songs on each list are from an artist already on that list.
In one week in March, Sweden’s Top 200 chart included 138 different artists, while Japan’s Top 200 chart was composed of only 88 artists’ songs. Surprisingly, USA’s Top 200 chart included 119 distinct artists; in our anecdotal perception of the music industry, we had hypothesized that this number would be lower, with certain artists “dominating the charts”.
Streaming ratios
Another way to look at disparity within the charts is to see the variation between streams of songs within the top 200. We can do this by comparing streams of a certain chart placement (#10, #50, etc) to the #1 song in their respective country. Countries with a higher ratio would have less variation in streaming numbers, and countries with a lower ratio have more variation.
Looking at the top #10 song compared to the #1 song, the differences are mostly uniform across countries, with a couple exceptions.
Moving down the chart to look at the #50 song compared to #1, the differences begin to get more pronounced between countries.
Comparing the #200 song to the #1 song yields the highest level of variation:
In March 2021, the USA, Turkey, Poland and Mexico had the least disparity in streams between the #1 song and the #200 song, while Japan, Morocco, and Egypt had the most. In the first set of countries, tracks in the top 200 share a more similar level of attention overall. In the second group, overall attention (measured in streams) is focused on a smaller set of tracks. For example, in Japan’s case, there seems to be a noticeable drop off after the 50 most popular songs. In Morocco, the popularity of the #10 song already starts to drop off at a faster rate than in most other countries. Egypt’s data stood out most in this analysis: The top song, “Ghaba” by Marwan Pablo, was extremely popular compared to every other song in the sampled month, having 4 times more streams than even the #2 song!
Conclusion
The number of unique artists has an interesting correlation to streaming variation within these Top 200 charts. The three countries with the lowest number of unique artists — Japan, Morocco, and Egypt — also had the highest disparity in streams between the highest ranked song and the lowest ranked. Conversely, the countries with the least disparity in streams also generally have the most unique artists within the top 200 — Including Turkey, the USA, and Poland. Despite anecdotal perceptions about certain music “dominating the charts” in the U.S., it appears that the U.S. is one of the less exclusive countries represented in the data according to our metrics.
There are many other questions that could be investigated using Spotify’s data. One factor not addressed in-depth in this article is the global nature of the music industry. As previously mentioned, there are three songs that appear in the top 200 charts of all nine countries. All three songs are by American artists, and measuring the overlaps between charts could give insight into the extent that countries like the U.S. influence the rest of the world’s popular music. Additionally, with extra columns indicating the language of each song, we would be able to measure phenomena like the number of English songs that appear in the charts of countries where English is not the primary language. There’s a lot left to be explored in this data.
