In the last decade Spotify has not only changed the way we consume music, but has been able to alter the music industry itself. It is the most popular streaming music platform, which is why the data it releases are very significant to take the pulse of the music business.

But how does Spotify measure the popularity of an artist? Who are the best sellers and how far do they separate them in terms of popularity from the bulk of artists supporting the music industry? What conclusions can we draw from an analysis of these characteristics? A study on the popularity and followers of artists through Spotify and the relationship that exists between them can provide us with valuable information in this regard. Our data analysts shed a little light on this question, helping us to understand this entire universe a little better.

The study, carried out in April, reveals that Spotify has 1,320,597 unique artists, of which the five with the greatest number of followers are Ed Sheeran, Drake, Ariana Grande, Rihanna and Eminem, as can be seen in the following table:

La imagen tiene un atributo ALT vacío; su nombre de archivo es image.png

 

 

Artists by number of followers

How are artists distributed according to the impact they have on Spotify? We will put our attention on the following graph:

La imagen tiene un atributo ALT vacío; su nombre de archivo es Numero-seguidores-Spotify.jpg

The horizontal axis represents the position that artists are in according to the number of followers they have on Spotify, from 0 to 100. It is striking how the number of followers that 40% of the artists with the least reach have almost the same number of followers of the artists who are in position 80. The difference is imperceptible. The best sellers, meanwhile, represent 1% and are the ones with the most followers. Between these and the rest of the artists we see an average abysmal distance.

We will use a logarithmic scale so that the difference is not so pronounced and so we can get a better idea. In the following graph, being one point higher on the vertical axis means multiplying by 2.7 (almost tripling) the number of followers compared to the previous point. We see how the bulk of the artists who are on Spotify, represented by the horizontal axis, are between 3 and 10 of the scale, with the number exceeding point 11 being minimal:

La imagen tiene un atributo ALT vacío; su nombre de archivo es escala-logarítmica.jpg

 

 

Artists by popularity

Let’s look at the graphics above but focusing our attention on the “popularity” that Spotify gives artists rather than the number of followers. This “popularity” is an internal Spotify variable that they calculate giving them a range from 0 to 100.

La imagen tiene un atributo ALT vacío; su nombre de archivo es Gráfica-popularidad-1.png

“Popularity” is an indicator that Spotify calculates. It is designed and adapted to have this shape, which works better than the number of raw followers. The curve of the graph is so regular, and therefore artificial, for this very reason: it is an adaptation made by the platform itself so that the jumps between the most successful and least successful artists are smoother and more calibrated than the data of followers real.

This graph of popularity on a logarithmic scale would look like this:

La imagen tiene un atributo ALT vacío; su nombre de archivo es Popularidad-logarítmica.jpg

 

 

Followers vs Popularity

Spotify’s “popularity” and the number of followers of the artists are not as directly related as it may seem. The separate charts indicate that “popularity” is a softened version of the number of followers, because the jump in the number of followers among the artists who are most famous is very large compared to the rest. But what we have seen are properties that are compliant with large-scale data. The particular behavior of each artist is more complicated.

In the following graphic each circle is an artist. The vertical axis indicates their number of followers and, the horizontal axis, the popularity value that Spotify gives them:

La imagen tiene un atributo ALT vacío; su nombre de archivo es Popularidad-vs.-Followers.jpg

It is appreciated that the general behavior reverts to what we have already seen: the most popular artists multiply the number of followers of the other talents several times.

Analyzing this graph, it is not only remarkable that the artist with the most followers (Ed Sheeran) has more than a million compared to the second (Drake), but that among the artists who have 0 popularity, there are several that have a very remarkable number of followers. . This may be because, in the time it takes for Spotify to update its internal popularity indicator, these artists are new and have a very significant growth.

 

 

Growth

It is interesting to look at the growth charts to understand how the bulk of artists behave. For starters, the graph of the difference in the number of followers (in absolute terms) has almost the same shape as the graph of the number of raw followers of the artists we saw at the beginning, as expected:

La imagen tiene un atributo ALT vacío; su nombre de archivo es Crecimiento-artistas.jpg

The 1% continues to dominate the growth market in absolute terms, and there is another 1% of artists that are losing followers. It is striking that only 2.5% of artists (more or less) gain more than 1,000 followers during the 3-4 weeks in which the data was collected. This indicates that, in absolute terms (which are the main indicator) more than 95% of artists remain stable.

But in the following graph we can verify that the extremes can actually vary quite a bit (there are artists who go up more than 40 popularity points, and others who lose more than 50). Among these cases, the number of followers does not have to vary much. In order to gain a better interpretability of why this happens, a more specific study would have to be carried out.

La imagen tiene un atributo ALT vacío; su nombre de archivo es crecimiento-artistas-abril.jpg

To finish, we consider the graph that shows the difference in followers “normalized” by the average of the measurements. It should be noted that it closely resembles the shape of this graph that we have just seen, but in this case we do have a really interesting variability between the number of followers on Spotify:

La imagen tiene un atributo ALT vacío; su nombre de archivo es Crecimiento-normalizado.jpg

In this case we can draw at least two very interesting conclusions:

  • The truly remarkable group of artists with relative growth is around 0.5% of total artists.
  • There are 20% of artists that decrease or remain without growth. The logarithm of 0 or null values does not exist, so the first percentiles do not appear on this graph.