One of the things that seems to distinguish ‘classical’ from ‘popular’ music is the fact that the same classical composers and works can remain at the top for very long periods of time – decades, even centuries – whereas popular music songs and artists can reach the top of the charts, sell millions of records, and disappear within a matter of months. But is this difference real?
The obvious problem is that we are not comparing like with like. We are comparing classical works and composers with popular music performing artists and their recordings of particular songs. The classical world tends to focus on composers and works, but in popular music these are much less important than performers and their interpretations of works.
For classical music, we could look at, for example, the most popular works or composers performed at the BBC Proms – and we would find many of the big names in the years 1900-1910 also appearing in 2000-2010. Beethoven, Mozart, Wagner, Bach would all be high on both lists, with many of the same works. But the performers would be different (although some orchestras might appear in both periods), and the same works would have been performed differently: Bach’s orchestral works, for example, would usually be performed nowadays in an ‘authentic’ format, more-or-less as Bach wrote them, whereas a century ago the fashion was for more ‘romantic’ arrangements, with a fuller sound, more instruments and even added textures.
For popular music we could look at the singles charts and note that there is virtually nothing in common between two charts six months apart. But the top composers might still be represented (think of Burt Bacharach, Bob Dylan or Lennon & McCartney during the 1960s-70s), and the same song might reappear many times in different cover versions by various artists.
The pop charts are a measure of weekly record sales, indicating a particular measure of short term success, but they do not capture other measures of success, such as how many times those records are listened to, the number of plays on broadcast media, the number of tickets sold for live gigs, or the various social media metrics that have emerged in the last few years.
Classical concert data, on the other hand, measures success differently. Concert programmes may be selected by promoters to bring in audiences, to show off a particular soloist, or to minimise the financial risk of the concert. They do not necessarily reflect alternative measures of success such as record or sheet music sales, performances by amateur orchestras, positive critical reviews, or what people choose to listen to at home.
This is an example of where our perceptions are at risk of being shaped by the data that happens to be available. The classical and popular music markets work in quite different ways, and the available information is driven by those differences.
It is hard to think of a source of data that would enable us to compare classical concert popularity with something equivalent in popular music. However, the long touring careers of some of the top rock and pop performers from the last sixty years, who continue to play their old songs several decades after they were originally hits, suggest that, perhaps, the contrast with classical music is not as great as we might have thought.
In recent years, separate record charts have been compiled for different musical genres, so it is now possible to compare classical and popular music on this basis. The website www.officialcharts.com contains a wealth of data on the UK record charts going back many years. Using this data, and taking the top 50 album charts for ‘classical artists’ and ‘all albums’ for the (arbitrary) five-year period 2000-2004, it is possible to measure the correlation between charts several weeks apart, resulting in the following picture.
For this analysis, the non-classical album chart was derived by removing the classical albums from the top-100 all albums chart, and renumbering accordingly. Each dot represents one pair of weekly charts (blue for classical, red for non-classical), plotted according to the number of weeks separating them (on the horizontal axis) and the correlation between them (on the vertical axis). The lines show the average correlations by time lag.1
Although the correlations start off in week 1 at about the same level, the classical album chart shows noticeably higher lagged correlation throughout than the non-classical chart. It takes about 11 weeks for the classical chart correlation to fall to 50%, and even after 50 weeks the correlation is still almost 25% on average. For non-classical albums, 50% correlation is reached in just 6 weeks, and by 50 weeks the correlation is more or less zero. What this tells us is that the classical chart changes more slowly than the non-classical chart, and that successful classical artists are more likely to enjoy a long run near the top than their non-classical counterparts.
So this does support our original hunch that success in classical music is longer lasting, slower moving and less transient than success in the world of popular music.
However, the effect is nowhere near as great as in the examples discussed at the start of this article. If we carry out the same analysis for the UK singles chart, the correlation coefficient falls to 50% in three weeks and is close to zero after about 15 weeks. Repeating the analysis for composers’ appearances at the Proms, it takes over 60 years for the correlation to fall to 50%, and it never gets below 25%. So the differences between classical composers/concert performances and classical artists/album sales, or between popular singles sales and popular album sales, are much greater than the difference between classical and non-classical album sales.
We are constrained by the lack of properly comparable data, but I think we can conclude from this analysis that there are differences in the patterns of success between classical and popular music, but that they are primarily a result of the different ways in which these two markets work and the ways that success is measured in them.
- The correlation was calculated between the reversed chart positions in the respective weeks for all albums appearing during the five year period. So number 1 in each week scored 50, number 50 scored 1, and all albums not appearing that week scored zero. The explanation of correlation on Wikipedia may be useful if you are unfamiliar with it.