We Need To Talk About Streaming (again)

Last night I participated in a Music Tank seminar on streaming music.  It was a vibrant and valuable debate with a healthy diversity of opinion.  Below are brief highlights of my opening keynote, including some exclusive data from record labels and from Spotify.

Streaming isn’t the paradigm shift, increased convenience of music access is

Streaming is no new thing.  Napster, Rhapsody, YouTube have been with us for many years.  What changed is that Spotify made it work with elegant simplicity, wrapped up in a consumer-friendly value proposition.  Of course Spotify had timing on its side too, coming to market once most of us already had broadband and at a time when a rapidly growing share of us were getting smartphones with data plans.  And of course timing is everything in business.

Timing aside though, we should be careful not to get hung up on the idea of streaming as an alternative format to the download.  It is not.  It is simply a different delivery mechanism for digital music, and when you factor in cached streams the distinction blurs further.  Streaming versus downloading is tech speak.  All music fans are interested in is being able to listen to the music they want, when they want, where they want.

Rebooting the conversation

Streaming music, and Spotify in particular, has been cause of much controversy and debate of late.  I’ll come on to some of the causes later but it is first worth taking stock of what we actually do and don’t know about streaming.

• What we know. Streaming is proving popular with consumers at a time when download growth is slowing. But many artists are not fully comfortable with the model and feel that they don’t get a fair enough deal.  A dynamic which is complicated by the fact there are many different types of artist deals.  Scale is key to streaming being successful (you don’t make money off dozens or hundreds of streams).
• What we don’t know.  We don’t know yet whether streaming cannibalizes sales.  Whatever data you see on either side of the argument we are simply too early in the evolution of streaming to draw conclusions.  There simply isn’t enough empirical data.  We need a few more years yet and even then separating cause from effect is challenging at best.
• What we suspect.  It is looking like streaming does help reduce the amount people use file sharing.  Again, the evidence isn’t definitive and there certainly isn’t sufficient evidence to suggest that the number of people using P2P etc is declining due to streaming, but intensity of usage perhaps.  Smaller artists don’t seem to do that well out of streaming.

 

Access based services are the first post-transition technology products

Any new technology looks more like what came before than what will come next.  After all we only have the past and present as our reference points. Thus when a new set of technologies emerge they begin with transition technologies.  The first car was a steam powered horse-less carriage (see figure 2).  It was a transition to the first internal combustion engine vehicle and it wasn’t until the 1950’s that we really started to see automobile form factors that had fully thrown off the horse drawn carriage heritage.  Digital music is no different.  The download was the steam powered horse-less carriage, a really useful transition tool to help us bridge the gap between analogue and digital, but just that.  Access based services are the first steps towards the internal combustion engine, services that leverage some of the unique capabilities digital presents, rather than just using the web as a delivery mechanism.  But it is still very early days, we are not even at the Model-T yet.

Putting streaming income into context

A number of record labels provided Music Tank with data illustrating the level of income across various platforms which can see here at aggregate level (see figure 3). This chart uses the income from a download as a base of 1 and then income from other sources as a multiple thereof, shown for labels and for artists.  Note that the artist data is 3^rd party licensing income only and does not reflect songwriter income etc.   The data suggests that an artist requires 80 streams to match the income from one download.  However data from artists suggests it is more than 200 streams.  And this rate varies massively depending on the nature of the deal an artist has struck with their label (e.g. whether they are paid on a share of net income basis or on points) and what share intermediaries such as distributors take.  It is also impacted by what deal the label has struck with a service.   One smaller label claims, somewhat dubiously, that the rate for them is closer to 2,000 times.  Whatever the exact rate (and there isn’t just one) you have to  stream a lot of music to get the same income as a download, but much, much less than a web radio stream or radio listens.  It take more than 5,500 national BBC radio listeners to generate the same income as one download in the UK

The labels’ take on streaming

Some record labels also provided Music Tank with some of their views on streaming and how they see it in the bigger revenue picture. Quickly summarized these are:

• Markets with a strong streaming sector also often have stronger overall digital growth
• Streaming is now growing more quickly than downloads
• Streaming can be 50% of an artist’s digital revenue in some markets
• Streaming consumers and download buyers do not strongly overlap
• Streaming subscriber ARPU is often higher than download buyer ARPU

Spotify’s take on streaming

Spotify also put some data on the table (see figure 4) showing how a major global artist’s catalogue fared following the release of their album the same day to stores as to streaming.  Obviously this data is positioned in the context of the cannibalization and ‘windowing’ debates (which I’ve contributed to here). The data doesn’t prove anything either way in terms of cannibalization (i.e. it could be interpreted as streaming activity does well when an album does well or it could also be viewed as lost buyer activity).  However it does make a compelling case for the degree to which an artist’s back catalogue can be significantly boosted on streaming following an album release. There are some well voiced concerns that streaming favours big name artists, the head rather than the long tail, but if it does then it appears to do a good job of mining the long tail of the head!

The potential of Spotify’s Developer API strategy: an API for Music?

In the last 6 months digital music has two developments of potentially seismic proportions that through their subtle brilliance many haven’t yet appreciated their actual importance.  One was Facebook’s content dashboard strategy.  The other was Spotify’s Developer API.  Of course APIs are no new thing, but if Spotify can reach a hundred million plus total users then its API has the potential of becoming a de facto API for music.   Allowing developers to skip seeking licenses from rights owners and using Spotify’s instead.  It is a crucially well timed move, coming just as investors are turning away from investing in services that require licenses (you may have noticed by now that impecable timing is one of Spotify’s strengths).  Investors have tired of funding license advances for services that often, as in the case of Beyond Oblivion, don’t even make it to market.  The labels still get their digital income but investors are left with a debt write off.  Index’s highly influential Saul Klein went as far as stating that he won’t even invest in  start-ups that require rights owner licenses.

Making the right comparisons

Crucial to the streaming debate is making the right comparisons.

• Streaming does not = a download
• Streaming does not = radio
• But Streaming does = (download + radio) ÷ ??

The exact balance is in flux but the conversation must recognize that a direct comparison with either is off the mark.  What we don’t yet know, and won’t for a couple of years, is whether streaming is pulling its users from green field and thus growing the market in a truly additive manner, or whether it is instead catalysing the organic digital transition, converting those consumers who would have gone digital anyway.  If it is the latter then questions about the income from streaming users compared to other digital customers becomes a more pressing one.  If it is the former then it frees us up to look at the scale picture with fewer reservations.  If these customers simply weren’t ever going to adopt a different digital service then we can start to discuss how low we can bring pricing to drive even great numbers.  The elephant in the room is that £/$10 is just too much for mainstream consumers.  It needs to be close to £/$5 to really break into the mainstream.  And you can only make that business case with genuine scale.

Conclusions

• It is too early to make conclusive judgements about streaming affecting sales or piracy in the near-to-mid term
• Long-term, music consumption will shift from ownership to access
• The streaming debate is clouded by conflicting artist statistics and concerns
• More artists need to be better sold the story by labels and by the services themselves, and some deals may even need revisiting.  Greater transparency is key and record labels have a big role to play here – there’s only so much services can do themselves
• Streaming is neither a radio replacement or a download replacement, it has some of the best of both

As for the legacy of streaming?  Streaming will help make Facebook the most important player in the digital music market by 2013.