ratpie

Universal Music has been turning in some very unimpressive numbers for parent Vivendi over the last few years. Traditionally record companies blame everyone other than themselves for their failures (I did a brief survey some years ago here). Now in the frame is Spotify, which relies heavily on drawing users in on its free offering, from where it can sell data about them to advertisers and hope they upgrade to a paid plan. That free tier, said an unnamed Universal Music executive to the Financial Times, is not working hard enough for the music company, and moreover it steals consumers who might otherwise be buying downloads. iTunes, by far the dominant download store, is suffering slow but steady attrition.

Quoted in the same Financial Times article (March 2015), Spotify managed to both prove and deny this simple proposition:

The company also rejects the suggestion that its ad-supported tier is responsible for declines in download sales. “Spotify is not cannibalising iTunes,” it says. “Spotify is monetising people who have never been monetised before.” Only 12 per cent of former iTunes users are on Spotify and of that number more than 40 per cent subscribe to the paid tier, it adds.

You don’t need PhD level statistics to see that if 1 in 8 of the people who stop using iTunes do so because they are spending more than they would on a Spotify subscription, and if nearly half of those high rolling stoppers trade down to a spend which is capped at £120 per year including VAT there’s going to be less money in the system for recorded music. Access to all the music they either lacked the time or the money to acquire might convince some lesser spenders to push the boat out, but for the majority of infrequent or parsimonious music buyers, big value 100 track compilations would remain a better deal. The gifting market has a bit of inertia and is more likely to stick with CDs, or default to an iTunes voucher.

More contentious however is that free of charge tier in the subscription service model. For the many who never or only very rarely bought music in any form, the subscription business is as irrelevant as hipster vinyl. But again common sense says that there will be  cannibalisation; between free tiers of service and whatever advertising or taxpayer funded licence fees can be scraped together for broadcast or webcast services, and between free and the paid tiers. And while some services, most notably Spotify and its direct competitors such as Rdio and Deezer, strive to find a balance between free and paid plans, for others, including YouTube and Pandora, paid plans remain commercially insignificant.

I previously calculated very approximately a net loss to recorded music revenue of about 30p for every £1 of subscription revenue. Some analysts say that streaming is just too expensive, and that the revenue maximising full service price is more like £3 per month rather than £10. One of the clearest explanations, by David Touve, can be found here and in more detail in this presentation here (but please ignore the Swedish Internet traffic data as there was a much more plausible explanation than iPred for the drop recorded). At such low prices the cannibalisation effect would be even harder to deny. And even within a single service it seems hugely important just what you get for nothing and what you have to pay for.

So this has been a brief survey of just some of the moving parts in the market for recorded music. But why is it that spats over the finer points of music service revenue models reach the world’s financial press? And why do some of the biggest and most savvy businesses seem to be pushing so hard for apparently sub-optimal results?

That cannibalisation word is a horrible one and nobody is eating people we hope – the idea is contention for finite resources. Record companies and services each have their specialised capabilities that combine to capture value from music, so to some extent they must cooperate. But naturally all music services compete with each other, as do all record labels; and in the middle is a pot of value that could legitimately go to any of the contestants, or indeed to consumers, or to the providers of the devices and networks that connect us to the music.

In that pot is the explanation for the more inflammatory proposition of my headline, that Universal cannibalises music. Here’s how Vivendi describes UMG for investors:

Undisputed global leadership position in recorded music:

• More than 30% market share

Translate that into a set of rules governing what models and prices a dominant record label would be willing to tolerate, and the trajectory of the market becomes obvious. Dominance means that a market participant not only wins, but is also able to choose the terms of engagement such they will always win. In music that means Universal takes market share away from all the other record companies, and to the extent that it can, also takes value from music services. Those services in turn need to see more of the value they capture also coming from the other labels, and the biggest squeeze will inevitably be put on the smallest. The digital music market has the same kind of inevitability to it as a modulated Conway’s Game of Life. It ends, by the way, when there is either nothing left to contend for, or nobody left to contend for it.

And the lesson I guess for anyone who makes or owns music that they believe is as valuable, or more so, than Universal’s, is not to swim in a shark pool. It’s probably time to test the always dubious notion that a a platform without chart hits (dominated by UMG and Sony, with Warner and the indies taking turns every few weeks to get a look in) would be unacceptable to a consumer with more developed tastes.

Big record labels are gearing up to move their release scheduling so that all new records come out on a Friday at a minute past midnight. They argue that everyone knows anyway through social media when a record is released; that record stores for those buying physical products need to know what day to rotate their displays (they do now of course, but risk being out of sync with download and streaming release dates); that new music should be introduced when interest in music is at its peak in order to compete better with other entertainment options. And also, that simpler planning means better coordination for global marketing campaigns, and also for their flip-side, anti-piracy activity.

This all looks convincing. Some High Street shops might prefer a different day (Fridays must already be busy for non specialists as everyone gets ready for weekend shoppers) but will welcome being at less of a disadvantage to digital retail and streaming. Much of the rest looks like simple common sense.

But let’s look at some of the assumptions here. For a start, this is clearly built to fit the Western Christian working week. OK, so the major music markets might work on Western weekly cycles, but at the very least this shows some overreach from record label representatives in the UK & USA who led the decision to move to Fridays. So much for ‘think global, act local’. Also underlying this move is the idea that new music discovery is a short burst of highly competitive attention. It allows no time for second listens, or getting used to an unfamiliar style over the week before deciding to spend your money and make your mark on the chart.

But are we not increasingly ignoring charts, and making our own journeys of discovery through social media and infinite libraries of music? That would be nice to think, but the money says otherwise. Those charts are important, not just as a signal of success, but also as a vector by which music can move from a niche audience and find broader support. So constricting the gateway is an inherently conservative strategy, favouring the bigger and better resourced marketing teams, and familiar artists. It would not be too far fetched to suspect that the same voices calling for a global release day will soon be trying to coordinate global charts, all based on the ability to mobilise marketing clout in a very compressed and highly contended window of essentially two day’s sales.

This is a kind of market shrinking madness, inevitably benefiting bigger companies and those with access to TV, which is probably no accident. The recorded music market is essentially static, so if the dominant labels want to improve their financial performance they need more market concentration, and less cost.

For the rest however, young artists, independent record labels, and anyone who wants to roam outside of the mainstream music playpen, we need to get a bit more organised and find ways to support truly local music and music retailers, outside the hypercompetitive sharkpool that global release day will become. May I start by suggesting Folk Monday’s, Welsh Tuesday’s, Opposite Latitude Wednesday’s, Over 50’s Thursdays…. you get the picture. And then we will be fortified against the torrent of Global Pop Friday.

Some ‘state of the nation’ observations about where we seem to be at the end of 2014 in the world of digital music:

1. Only an extreme optimist would make a meaningful bet on the recorded music industry as a whole growing significantly in the next few years. Sales of downloads are likely to show slow growth or stasis, streaming will grow, some new territories will emerge, but what growth there is will be about enough to offset the continued slide in physical product. On demand playback, over the air, over wires, or from encrypted caches on devices and hard drives, is getting more people habitually paying for music, but for many previous download and CD buyers it is a better value replacement for owning files and discs.

2. We have, finally, a reasonably well set up supply chain in digital music, with a collective path to future improvement through DDEX, which has changed in an important way by including compliance, and is also now touching ancillary areas of metadata management and communication. There is renewed impetus to address other shocking failures in metadata and identifiers. As a whole the industry is probably still investing, but a return on that investment is now firmly in sight.

3. Good as it is, and with proven scalability to support the 45 million tracks now commercially available, our metadata capability is very skinny, and is not even starting to approach the depth and richness of the LP sleeve or CD booklet. This is a serious consumer issue; if we don’t feed the passionate or even the merely curious we are devaluing all recorded music, and taking oxygen from artists and the industry. What commercial databases there are are an insult to fans with their bad prose and out of date reviews and biographies. One square artwork image, and one artist photo, is equally poor. Record labels and services need to collaborate to improve this dramatically, or face a future where recorded music is a utility, rather than a passion and an obsession for fans.

4. The professional part of the industry needs to differentiate from hobbyists, and the best way to do that is by embracing every aspect of product quality, from the microphone right through to the press release. Twenty years ago it was OK to represent the product as an unfortunate impediment between the art and the audience. Now not so much; the world is rediscovering its love of authenticity and craft, and music needs to be part of that journey through recognised high product quality standards and extensive consumer education. This applies to digital as well as physical, where it is visible already in the re-emergence of the vinyl record. Our love affair with DIY culture is no reason to demolish all formal signs of professionalism. This is not an argument for closed shops and exclusion however. Consumers will embrace the effort and care that goes into a music product in the context of fair and open markets.

5. One school of thought sees music, particularly pop, as a grand illusion, or less charitably a charade. At its best the pop music spectacle remains utterly compelling, and great escapist fun. We are also, however, in an age of radical transparency, where every action, word, contract, and posture is minutely scrutinised. For the music industry, romance of the kind eulogised in the famous Thompson quote, is dead. This means, I believe, that we have a very difficult path to walk, between the excesses of the past (behaving badly and getting away with it) to a future founded on a different kind of passion, for high quality and ethical product and business practices. But nobody likes a pious fool, so we must do this without jettisoning our humour and ability to connect emotionally with music fans.

There is nothing special about the music industry; we see trends in business and in consumption across the world promoting ethics, quality, transparency, and a shared sense that we should use what we do each day to improve the world we live in. These thoughts for 2015 are offered with hope, and with goodwill to all who work in the music industry, and all who love music.

Continued debate about how to share revenue generated by on-demand music streaming services is less enlightening than it could be, mostly because it is conducted in an almost information free context. We don’t really know how much money is being brought in, how it splits between subscriptions and advertising, or what’s in the deals that govern the payments to the recording owners. Nor do we know what the contracts between the artists and the labels say, including what the artist traded off against what now seem to some to be grossly unfair royalty rates.

Different belief systems will of course have more impact on the opinions of the debaters than any evidence might; that is just human nature. One of the fundamentals is whether the price of something is considered to be the best available proxy for its value, or whether it is an ethical expression of some kind (look for the word ‘should’ to sift out the latter type). When lots of people collectively buy lots of things it is easy for the relative values in each of those things to get flattened, and a ‘fairness’ approach to pricing might even point in the direction of ‘one price fits all’. It could perhaps also be argued that each track in a catalogue has a potential range of values from highest to lowest; it just needs to find enough ears for this to emerge.

Common sense however suggests that some recordings will be generally valued higher by more people, and some might even be seen by most as a blight which warrants compensation for exposure to them. And by the time your ears tell you which is which it is too late. It is very hard to unhear something, and even Cher has not infallibly managed to turn back time.

When the consumer clicks the play button some noise must happen if the on-demand model is to become universally trusted and accepted. Any pricing or licensing scheme that could withhold a track at the moment of demand is therefore unlikely to work very well – driving away the subscriber is not how to create long term revenue growth for all. Search is also a sensitive moment for the consumer, and no service wants to be forced to advertise incompleteness.

Yet it seems clear that we need a more dynamic wholesale layer in the market, so that the reward and incentive is much better allocated than currently. Here are a very few simple suggestions, which might stem the criticism for long enough to find out if an on-demand world is capable of supporting professional recording artists making high quality recordings.

1. The first listen of each track by each listener could be royalty free, or even create a debit against the licensor’s account. This automatically shifts revenue towards artists who can make tracks listeners want to hear more than once. Preference data (favouriting, playlisting, caching) might then start to give some clues about likely future value, with relative rates adjusted accordingly.

2. Historical play counts could be used to stratify the catalogue, and each tier could be allocated a royalty pool. Models of what affects play counts will be naturally very complex, so this might be a relatively weak tool to discover how tracks are valued, but it could effectively, for instance, be used to incentivise external support for ‘growers’, or just ensure that providers of unpopular and unloved music had fewer reasons to market within a service.

3. Similarly, style and genre metadata could be used to differentiate between royalty pools, allowing a service to encourage and reward specialisation. This just reflects what happens in the downloads market with specialised stores. Combined with play count data it would also encourage better categorisation – faking a genre for a higher per play rate could quickly get punished.

4. Subtly different from (1), the first day, or week, of a track’s life on a service could bear a discounted royalty rate, pushing more revenue collectively towards the later stages of the sales envelop for the more popular tracks. This would explicitly drive windowing behaviour, which might end up being seen as overall positive for the market by supporting higher prices for downloads.

It would be surprising if ideas like this were not being actively considered all the time by services and by the more sophisticated music providers. While they stop short of providing the wholly dynamic wholesale marketplace that would inherently be capable of rewarding higher value with higher rewards, they have the advantage that the skills required to optimise strategies on either side are close to the marketing and release management skills that already exist in the industry. We don’t all have to become futures traders overnight. And underlying all of it is a drive to discover a sustainable way to incentivise new production of a great variety of new music recordings. That has to be a decent goal for all of us.

Spotify took to its blog October 2014 to let the world know that if the ants have megaphones, in Chris Anderson’s rather superior phrase, Spotify’s megaphone is bigger. A ‘curated playlist’ had driven a track up the charts in the world’s biggest music market. The track was Waves from Mr Probz, and it would be wrong to think it was a grassroots phenomenon. Before it was picked up by Spotify the song had Red Bull marketing euros behind it.

Meanwhile, in a nostalgic twitter trawl I was taking music tips from Victoria Beckham (I helped Virgin Records rule the Internet in the 90s and did several projects with the Spice Girls).

Love this soooo much!! Girl Power!!! "Bang Bang" by Jessie J, Arianna Grande and Nicki Minaj x 🎤🎼🎶 😉 x vb http://t.co/W1dckPfnjd

— Victoria Beckham (@victoriabeckham) August 8, 2014

A great track from the music industry’s aristocracy, including legendary producer Max Martin, and as Posh says, Girl Power!!!, along with prime time TV and wardrobe malfunction support. At 79 footnotes Bang Bang’s po-faced Wikipedia page narrowly beasts Waves (74 footnotes) but has much much more information. The fanzine and fansite obsessive compulsive fact hoarders have migrated to Wikipedia, including the confidently named MusicLover (who pays you???) who seems exclusively chart focused and has a prodigious output. Big data enthusiasts might like to consider a footnote count as a new input to their algorithms.

So Spotify is telling us that its ‘curated’ playlists can drive a track to the top of the chart in the US; meanwhile the top table in that market is demonstrating an ability to deliver success in all the usual ways.

And despite all the footnotes the information available about each those tracks is remarkably thin. Allmusic.com delivers production and performance credits at album level, but not at track level. So for instance we know that Nick Barr might have played violin on Bang Bang, but not categorically that he did, or why he played violin rather than his usual viola. PPL’s repertoire search delivers an ISRC, USUM71409737, but nothing more exciting than that. BMI delivers a work number, #17595803 and some of the composers’ real names.

So the narrative on neither track is particularly deep or interesting. Pop comes out of a hermetically sealed box, and then is delivered by starmakers to a grateful but passive audience; crumbs of information are collected by wiki pagemakers and compiled into semi-literate parodies of academic articles (since when have fanzines had footnotes?). In an age supposedly of data and transparency it looks like nothing has changed since Paul Weller sang ‘the public wants what the public gets’. Today there is perversely less information available about the music than there was when paper monthly magazines had 1500 word articles to fill, and record companies commissioned writers for CD booklets.

But there is a big change and that is the flow of data and information about the audience, to the new marketing companies. Way back in 2003 artist Angie Waller produced a book called ‘Data Mining the Amazon’, looking at the connections between politics and popular music as revealed by Amazon’s ‘also bought’ data. It is (update, August 2017, sadly not) still available; Waller said in an interview:

I was surprised that books about military battles and corporate takeovers pointed to the soothing CDs of Enya and Sarah Brightman

Now, with mobile and on demand music and information services, a click on a play button reveals not just your affinity with corporate takeovers, but also who you are, where you are, and as all the algorithms light up like a Christmas shopping street, what you might be persuaded to pay for next. Fancy going to a gig? Here’s Bandsintown to help you:

Bandsintown Amplified is a multi-screen advertising platform that connects brands with a passionate community of fans and artists through an exclusive network of 50+ publishers.

They were watching me check the accuracy of the Paul Weller lyric quoted above, and are now one of about 5,500 websites that have cookies or other locally stored data on my computer, and another set of touchpoints in their database about me.

So the story of a song is now the thousands or millions of stories of the people that the song helps the new digital marketing industry track and target, and the new starmakers are the businesses that can find focus and buyers for that data. Nobody knows how big this market can get, which is why a business such as SoundCloud can justify a $700m+ valuation on a $29m trading loss.

The music industry, as one of the key providers of the content around which that data is created, is at a fork in the road. It will not be credible to position music as an emotional bond between artist and fan if the real commercial purpose is to deliver a focused cloud of marketing data to brands. The product will with increasing sophistication be made to fit the profile. Alternatively, at least some of the music industry can get real about the artist and the product, and deliver music with craft skills on display and explained in full, and showing a strong commitment to quality and authenticity. I suspect that many artists might not be comfortable trying to do both at the same time.

The long drawn out crisis in the recorded music industry continues, and one of the much debated questions is whether people are subscribing to streaming services instead of buying downloads and CDs. If they are, the worry is that the amount of revenue and the way it is shared might damage the recorded music industry even more than either piracy, or unbundling, or the shift to downloads did.

To be clear, there is no crisis in the supply of music; the commercially available catalogue is somewhere around 35 million tracks and growing fast still. There is so much more music on sale than any market is really demanding that it has become pointless counting the tracks. Most music goes unheard and unbought these days, except by the creator and a few close friends and family. No, what is feared to be under threat is the big machine that hires studios, professional arrangers and musicians, songwriters, producers, makes glossy videos, and generally delivers a high quality mass pop product, or indeed an excellent example of a genre or niche.

So the complaint from artists, and to an extent from record labels, is that the way the money flows means that they won’t be able in the future to make the sorts of records that they would want to release; and that for many who could survive on modest record sales, streaming payouts make recorded music an unaffordable part of their career. For recording artists, then, rather than for live performers, the game might be up.

And it has become a bit of a commonplace among the music industry’s critics and commentators that these old timers are just wrong! Wrong to compare streaming with album or single sales, and wrong about the rates they think they should be getting.

Responses range from ‘sell something else then’, advised by music industry consultant Mark Mulligan:

If streaming is eating into sales then the obvious next step is to drive other spending from streaming music consumers.

http://musicindustryblog.wordpress.com/2014/09/22/the-three-things-streaming-needs-to-fix-next/

And David Byrne, who surely knows a thing or two about making a living as a musician offered this thought in an opinion piece for the Guardian:

In future, if artists have to rely almost exclusively on the income from these services, they’ll be out of work within a year.

http://www.theguardian.com/music/2013/oct/11/david-byrne-internet-content-world

So are these ‘old guys yelling at fast trains’ as Moby described Yorke? Rather than doing the customary forehead slap and writing off the plaintiffs as Luddites who need to be put out to pasture, consider the following. On demand streaming is a way to hear what you want when you want. That is a big win for the consumer, because previously the only way to do that was to buy the record. It bears repeating, because the ‘streaming is not downloading’ camp suffer a logical discontinuity; streaming on demand from a subscription service is a direct substitute for playing on demand from music you own. The second half of the argument, ‘because the listener would not otherwise have bought the music’ is clearly nonsense.

In default of any other way to hear a track on demand, an on demand stream substitutes for a track sale. Some consumers would have bought, some would not, and as music is surely fairly elastic, whether that music would have been bought depends mostly on price and convenience. Here I think the moaning musicians have an excellent point. Some numbers will make this very clear.

The substitute for an on demand stream might not be a track sale at $0.99, or even $0.49. Perhaps it might start to be 1 for 1 at $0.09, or even $0.009, but that is still much more than a normal per stream rate, reported variously between about $0.003 and $0.007 to the label, and a fraction of that to the artist depending on their contract.

Here is food for thought. If you never played any track twice, and if clicking the play button completed a transaction, at $0.009 you could get about 30 tracks per day for a month for a retail value of about $8.20. At the same play rate per day, any subsequent replays would simply reduce the monthly bill, and the effective ‘price per play’ equivalent by 1/plays.

Three months into a $9.99 monthly music spree at those rates, and assuming no replays, each punter would have a collection of 2,700 music tracks – more than academic studies have found on college music fans’ laptops. So, check your own music library, and look at the playcounts of the most listened to tracks. What used to be worth $9.99, or even $17.99 is now available, along with the other 35 million tracks, for $9.99 per month. But you have lost the connection you had with the artist that allowed you to put a few dollars in their specific pockets for the pleasure of hearing their music, and the business now dictates that the rest of your relatively mindless consumption reduces the earnings of the artists you care about.

We need much better models, which take into account long term changes in welfare for everyone involved in the creation, recording, and distribution of music, so that we can think more clearly about the consequences of what we do as music lovers, and within the industry. Perhaps our conclusion might still end up being that as a commercial art form the sound recording has had its day; perhaps though we will find a way to nourish and reward a greater variety and number of its practitioners in a way that makes them feel valued, rather than discarded.