[caractacus]

The data is hugely skewed because they're only looking at the top 200 each day. There are 50m tracks on streaming services and the popularity of the long tail is growing.

The 500 most streamed tracks in the US received just 10.7% of total audio streams in 2018, down from 14.6% in 2017. There were 36.3m different tracks streamed across the year, up from 33.2m in 2017.

https://cdn.mbw.44bytes.net/files/2019/01/BuzzAngle-Music-20...

[3xblah]

Cassette and vinyl sales are both up from 2017.

What strikes me about a report like this is that the report is comfortable comparing sales to "consumption".

Are these really the same?

Third parties do not necessarily know how many times someone cues up a track on their turntable or tape deck. (Unless today's turntables and tape decks are being engineered to "phone home".) Instead they measure sales of physical media containing copies.

On the other hand, with respect to so-called "streaming" the report says about 85% of this type of consumption is from subscriptions. But curiously the report does not measure subscription sales/renewals.

Maybe I missed something.

[saaaaaam]

Cassette and vinyl sales are both up - but still represent a tiny fraction of music sales - vinyl Is about 3% or 4% of sales in the UK.

As regards comparing consumption to sales: this is a fairly standard “equivalent” used in the music industry.

I’m unclear on what else you are saying. You say “so-called ‘streaming’” but why is it “so-called”? It’s pretty clear that the market has shifted to access over ownership or “so-called ‘streaming’”. Subscription streaming services are growing rapidly year-on-year. The stats are easy to find - perhaps the author didn’t feel it was necessary to quote numbers that are fairly widely discussed in mainstream and specialist media.

[chadmhorner]

Agreed we're only looking at the top 200! Tried to be upfront about that :)

[jakequade]

Why? You can't claim to have any real insights on such a vast platform when you're capturing such a tiny percentage.

[craftinator]

He actually can claim it, and does so. And that tiny percentage he captures, well that's called a "sample". When you take a sample from a larger population, it is safe to assume it represents some of the same properties as the population as a whole. Figuring out which properties are shared is one of the reasons humans aren't being replaced by machines entirely in the analysis department.

It is your job as the reader to invalidate some of his specific claims using sound logic and evidence. Which of his claims are you refuting?

[autokad]

wow, thats an amazing long tail. now I dont feel so bad not knowing a single one of the artists. I'm in the tail, just like everyone else