Something strange happened to the music business in the late 1990s. Until then, the hit single was everything. Ok, so it did not secure an artist the same kind of revenue as a multi-platinum album or a round-the-world tour, but it did signify entrance into the exclusive world of pop-stardom. It meant, if only for a moment, that you had the attention of the music buying public, in that instant – you were a star.
Throughout most of the 90s, hit singles were plodding along at a solid rate of knots. Whitney Houston scored over 3million sales for I Will Always Love You in 1992, Coolio’s Gangsta’s Paradise came close to those heights three years later, while Elton John topped 8million with a re-hash of Candle in the Wind in 1997. Then the wheels fell off. The best selling singles of 2001 and 2004 were purchased less than one million times between them. The physical single was over – the age of downloads had begun.
The contemporary picture is even more confused. As well as downloadable singles, streaming services must also be factored in if we are to understand true influence. While artists used to release two, three or maybe four or even five singles from an album, modern albums are viewed as collections of singles by the public, with individual tracks listened to at will.
When English singer-songwriter Ed Sheeran released his latest record, Divide (styled as ÷) in March 2017, all of the Most Streamed songs on Spotify on that day were from the record. At the same time, songs from the album were racking up YouTube views too, further skewing the situation.
We can see that there is data here. The question is, what can we do to make sense of this data? How can analytics help us sort the true hits from the near misses?
The Disparate Approach
One way is just to view each platform in isolation, calculating dominance across each. This should be simple enough. Each TV show or movie is watched a certain number of times on Netflix or Hulu, and each song is streamed a discrete number of times on Spotify, Deezer or YouTube, so, at the very least, we can work out a winner for each platform. Then we can assign different levels of significance to each platform, based on their cultural influence, collate all the data together and, hey presto, we have our hit.
Sounds good right? However, great ideas on paper don’t necessarily translate to great ideas in reality.
Parrot Analytics are data interpretation specialists who work with content streamed via Netflix to ascertain what is performing well and what is floundering. In March 2017, they turned their attentions to a show which has been generating a lot of noise in recent months, and not for the right reasons. That show is Iron Fist, the latest addition to the Marvel Comics multimedia empire.
After crunching the numbers, Parrot Analytics returned the result that, despite the negative publicity and critical trashings, Iron Fist was in fact a success. However, Parrot are not privy to the actual viewing data gathered by Netflix, so how have the firm reached this conclusion?
Well, Parrot use a concept called “demand expressions” to understand interest and buzz. Basically, if people are getting excited about a show, movie, or record on Twitter or across social media in general, Parrot Analytics know about it.
Parrot track these demand expressions over time, analyzing how interest wanes in the days, weeks, and months following the initial release. Through an indirect approach – i.e. analyzing reactions, as opposed to actual viewing figures – Parrot Analytics arrive at a fairly good metric for defining success.
But such numbers are always open to interpretation. Promotional activities, intertextuality with other cultural products (plenty of that in the Marvel Universe), and numerous other factors can all push internet ‘buzz’ in one direction or another. So, if we can’t even agree outright on the victor from one platform, what hope do we have of bringing all that data together?
What do analysts do when presented with a problem such as this one? We compromise, and we create industry standards of best practice based on the consensus view. This is how hits are defined in 2017.
In the US, the Billboard Charts have never been completely straightforward. An array of different charts, separated into different genres and styles, have led to segmentation and to artists battling it out for dominance within their respective categories. The digital music explosion add a couple of extra layers of complexity to the cake – first came paid digital downloads, then free streaming, then on demand streaming services – and nothing would ever be the same again.
Now we have situations that look like this. In the first week of April, 2017, Ed Sheeran is top of the downloads chart with Shape of You, while Luis Fonsi remains top of the YouTube free streaming chart, with Shape of You in second. Meanwhile, on the On Demand Music Service streaming chart, 21 of the rundown’s top 50 songs are by Canadian rapper, Drake, including both of the top 2.
For Billboard, this is enough. Cross-referencing and collation of data are not necessary here, as cultural influence is plainly displayed across the tops of the unified digital charts, and within each genre in the physical sales charts. Just like there can be more than one World Heavyweight Champion in boxing at any one time, there can be more than one dominant cultural producer, too.
In the UK, the Official Streaming Chart is slightly more regimented. The chart, launched in 2012, uses an accepted list of streaming sources, all of which must meet a certain level of criteria. These sources – including Spotify, Deezer and Napster – are then used to decide who has scored a hit record.
But arguing about which model is best will get us nowhere. Instead, we need to agree. The hits of the near future will be defined by a range of different data points within a certain set of parameters. It is up to us to reach a consensus on what those parameters should be, and to prevent the whole thing collapsing into anarchy.