30 May 2015

That Sony Spotify Contract

Two weeks ago, The Verge published a leaked contract between Spotify and Sony Music Entertainment, which detailed the US deal between those two parties. A lot has been written about that already, but here I will add a few more points that I haven't yet seen discussed.

The contracts contain various bits about IODA. At the time of the contract, IODA was partly owned by Sony. Since then, IODA has merged with The Orchard, and is now wholey-owned by Sony. Any indie labels that do their distribution through The Orchard, and wish to moan about this contract...: Be careful what you moan about!

The sections about IODA do seem reasonable though. Basically, IODA content will earn the same as original Sony content, but those earnings are treated separately and are excluded from advance recoupment. Sony simply doesn't want to have to pay IODA customers parts of their advances, and IODA (and their customers) will get paid, even if Sony hasn't recouped its advance.

The topic of whether Sony artists would see any benefit from the advances -whether recouped or unrecouped- has been debated at length, and Sony has publicly stated that they DO share such advances and breakage with the artists on the same basis as actual revenue, which is a good thing. For completeness: Warner has also publicly stated that they share such advances with artists, but Universal has yet to make any statement on this.

But the Sony/IODA link gets more interesting in light of the $9M of advertising credits that Sony received as part of the deal. I would expect that Sony's first point of use for those advertising credits is their own repertoire. When it comes to the option of re-selling the remaining advertising credits, it would be logical to think of IODA. Sony aren't in the business of selling online advertising, but that IS one of the things that IODA does. Are there any IODA-distributed labels who have bought Spotify advertising through IODA during 2011 or 2012, and would care to comment on the pricing thereof?

Whether Sony credited their artists with any of such advertising resale money is of course a totally separate point, which also comes after the question of whether -and how much- IODA paid Sony for those credits. We know details of neither of these, so I will leave that to others to investigate further.

Another interesting point is the 60% royalty rate that Sony receives. Spotify claim to pay out around 70% of their revenue to rightsholders, which is generally considered to be about 55% to labels and the rest to publishers. This 60% therefore raised a red flag here. But this contract is about the US and Canada only, and we all know that the publishing-side streaming royalties for the US are considerably lower than in the rest of the world due to those damned Consent Decrees and Rate Courts. I am curious about the rate Sony receives in other territories, but this contract does not necessarily imply anything on that matter.

The Verge made a rather weird comment about the per-stream minima in the contract. They made them sound somewhat irrelevant, because "these rates only come into play if the usage-based minimum exceeds the revenue sharing model". This is correct, but approached from the wrong direction: The minimum is fixed and the variation is in the amounts based on the sharing model. Sony will get paid on a revenue-share basis, unless that revenue share falls below the minimum, in which case they then get that per-stream minimum. Thus, Sony is protected from Spotify's potential failure to sell sufficient advertising, and Spotify is willing to give such protection. That's kind of fair on both sides, not? Furthermore, I have been reliably informed that all Spotify contracts with labels and aggregators contain such minima for the advertising-funded tiers. These minima do not give Sony an unfair advantage, except for perhaps the actual level of the minima, but I can't comment on how these compare to the minima for other content providers.

The most interesting part of the contract is how the revenue share is calculated. The contract shows that the revenue share is calculated separately for the different plans that Spotify offers. Actual earnings from the subscription tiers are therefore not brought down by any low per-stream rates for the advertising-funded tier. The free tier may bring down the average per-stream rate, but not the total. As I argued recently in my "Spotty Maths" post on this blog, the average is an irrelevant statistic. It's the total amont of money for the total mix that matters.

Finally, The Verge wrote about the MFN clause, and on the 15% on ad sales by third parties. But those comments were so dumb and ill-informed that I am not even going to get into these here. They may become the subject for another "Spotty Maths" post some other time.

16 May 2015

The Digital Pie Debate

I have been a member of the UK's MPA (Music Publishers Association) for quite a few years. Joining was one of the best things I have done for my business, because being an MPA member has given me access to a lot of information and contacts that I probably would not have had otherwise. I am so impressed and involved with the MPA, that I am standing as a candidate for the Board this year.

I'm not very impressed by comments made by the new interim CEO Jane Dyball as a panelist in "The Digital Pie Debate" at The Great Escape yesterday. You can read about her comments here:
http://www.musicweek.com/news/read/jane-dyball-there-is-not-enough-advertising-revenue-in-the-world-for-every-digital-service-to-launch-based-on-ad-revenue/061797.

Jane apparently "refuted the concept of a 'digital pie' to be shared out between publishers and labels". I don't get this. There IS a pie to be shared. That pie is 100% of subscriber revenue. It's been generally accepted that the service provider takes 30%. The other 70% of the pie is to be divided between labels and publishers. Whether the current division of that pie is fair is a different matter. The division varies between countries, which makes that a very long topic that I defer to address another time.

Jane continues by stating that "We are still very early into the digital market". I wonder where Jane has been recently, because we have been in the digital market for more than a decade. That does not qualify as "very early".

What annoys me most is this: "[...] we're not yet at the stage where you can easily track money coming out of a digital service and going into a royalties statement". The services, aggregators and labels are all capable of dealing with all of this. Various commercial companies that deal with online mechanicals in the US also manage to deal with it. Those are all companies that are legally and/or contractually obliged to deal with every fraction of a penny, and manage to do so, whilst the PROs and MROs (societies that rightsholders join voluntarily) all suddenly complain about how complex it all is, and do half-baked jobs. Some are better than others, but PRS for Music are among the worst offenders on that front. They generally blame "bad data", which is indeed a problem, but there are a number of other more significant problems in PRSfM's processing of data from digital services. (Another long topic that I may address another time).

She also mentions the animosity between publishers and labels. For Musiqware, close to 100% of our publishing revenue comes from recorded music. There is a tiny slice of money for live performances, but those are usually by an artist who has recorded a particular song, and is thus also related to the recordings. Even our synchronisation revenue comes from recorded music, because we place existing commercially-released tracks into TV, film, adverts, and so on. The labels are the organisations that generate revenue for us publishers. We should treat them as our friends and allies. The only revenue we have that is not related to recorded music, is sheet music. And just how much money do we all make from sheet music these days??


But these comments distract from what may have been her core message: I do agree that ad-funded revenue on its own is not going to sustain the music business. Fortunately, the majority of streaming services are freemium, meaning that they have an ad-funded free tier to atrract new users, whilst most of the income they generate is from premium subscriptions. Notable exceptions here are YouTube and Pandora. The subscription-focussed mixed models of Spotify, Deezer, Rdio, Rhapsody all generate good income for the music industry, including the publishers.

As such, advances from new services could be an option, but I believe that minimum per-stream values are a much better option. The ad-funded tier is a marketing strategy, and we must make sure that the music industry does not pay for such marketing costs. Per-stream minima solve that problem, and the label-side contracts with streaming services do include such minima. I hope the publishing-side contracts also include these.

13 May 2015

Spotty Maths #1: Freemium

My friend Mark owns 50 CDs. He's gone totally digital now, and is a bit short on cash, so he has decided to sell his CDs in order to buy a new phone. Ten of his CDs are really good and somewhat rare, so he wants to sell those for £12 each. The other 40 CDs are less valuable, so he thinks he will only be able to get £2 for each of those.

Q: Should he only sell the somewhat-rare ones, or should he sell them all?

If he were to only sell the somewhat-rare CDs, he would make 10 x £12 = £120. That's an average of £12 per CD.

If he were to sell them all, he would make an additional 40 x £2 = £80, bringing the total up to £200. But adding these cheaper ones to the mix, results in much lower average of only £4 per CD (£200 / 50).

I would advise Mark to sell them all, because he would end up with more money to spend on his new phone.

But if somebody were to ignore the quantities, actual prices, and total amounts, and only look at the per-CD price, then that person would incorrectly conclude that it's better to only sell the rarer ones, because "that results in a higher per-CD price". But that is simply a bad application of maths.


To apply these maths to streaming services, replace the rare CDs with "premium subscribers", and the others with "free tier users" (1). Offering a premium service results in a certain amount of revenue at a certain per-stream rate. Adding a free tier to that, increases the total revenue, but reduces the average per-stream rate.

The anti-freemium lobby seems to look mostly at the average per-unit revenue. They argue that freemium is bad because freemium services tend to have lower per-stream rates than services without a free tier. But that is not all that matters.

What really matters in this debate is: What would those free tier users have done if there hadn't been a free tier? Nobody knows for sure, because there has always been a free tier. A small group of them would just subscribe to the premium service. But most of them would simply opt for other free services such as YouTube and piracy, which pay even less, or nothing at all, reducing the total income for the music industry.

If the music industry wants to see growth again, they will need to allow freemium models. Those arguing against it are either bad at maths, or may have some hidden agenda.


(1) Freemium users don't pay anything, but they do generate revenue through the advertising they get thrown at them.

1 May 2015

What makes a proper publisher?

In a recent committee meeting with fellow publishers, I was accused of not being a "proper" publisher. The accuser said this in response to my comment that I was not particularly interested in "securing Swedish covers of our songs", but far more interested in making sure to get paid for airplay on Swedish radio and mechanicals from online services in Sweden. (This exchange was in a discussion specifically about Sweden, but applies equally to other territories).

It highlights how not all publishers are the same.

The publisher who made the offensive remark, is a very traditional publisher. He works with songwriters who provide him with songs, which he then pitches to artists and labels to get recorded and released. That is a perfectly valid business model, but not the one and only "proper" publishing model.

I work with artists who write their own music, most of whom are DJ/Producers. Publishing is one of a number of services that I offer them through my company Musiqware. Those services are not about securing cover recordings, but much more about securing payment for all uses of their music, wherever in the world such uses occur. We make sure that their music is registered with numerous collecting societies (and various other organisations that license direct), and that those registrations contain as much as possible additional data to assist with identifying all uses of the tracks. It also involves sending all music to music recognition technology companies to ensure that uses are reported. We also pitch and license music to film, TV shows, commercials, corporate presentations, online videos, and so on, either direct or through third party agencies.

But we do not look for covers, because that is mostly irrelevant to the artists we work with.

Not all publishers offer the same services, nor do they all have the same priorities. There is more than one "proper" publishing model.