I’m a jazz guy, and although I spend a fair amount of time in the vaults of Verve, Riverside, Columbia, and many others, my first love is Blue Note.  The relationship started when I heard Stanley Turrentine’s Up at Minton’s in the early 90s. Stanley himself is impressive, but Up at Minton’s turned me on to the reason that I picked up the guitar – Stanley’s sideman on the recording, Grant Green.  Green was a session player on Stanley’s record, and after seeing his name on the back cover of the record, I dove down deep into his catalog, which turned me onto a wealth of other players, like “Baby Face” Willette, Big John Patton, Hank Mobley, Lou Donaldson and many more. It all started with Stanley, but the collaborative nature of jazz, coupled with the sideman information on the back of all the records, enabled me to dig in deep to a catalog that rarely disappoints.

Back in the day, this discovery process was mostly a physical endeavor.  I would take the Red Line to Harvard Square, and make a loop through all the areas record stores, starting at Mystery Train, then onto Second Coming, over to Looney Tunes, Tower, and finishing up at Newbury Comics. Excellent way to spend a Sunday.

Of course, physical retail has constricted since then, and while there are a lot of positive attributes associated with the convenience of digital music, one thing that has not been replicated particularly well online has been the discovery process.  I have trusted sources and curators that help, but there are few online options that replicate the experience of walking into a store like Mystery Train, finding an amazing record with a little known sideman, and heading down the path of discovery for a brand new artist.  This is the fundamental reason I like the Blue Note Spotify app so much.  It is the best option for replicating this discovery process that I loved so much back in the day.

Walter Gross is the Senior Director of Digital Marketing at EMI, as well as a 1993 Berklee Grad in Jazz Composition.  Walter spearheaded the development and execution of the Blue Note app on Spotify, and I spoke to him a couple weeks back about the creation process, what he was trying to accomplish with the app, and the opportunities that Spotify’s API offers technologists.


===

Mike King: What did you want to accomplish with the Blue Note Spotify app?

Walter Gross: There was a small group at EMI that wanted to give this app a try, including myself.  I’m a Blue Note freak, and I remember being in a Berklee course with Jackie Beard [Professor in the Woodwinds dept] in 1994, and listening to a Joe Henderson Blue Note record, and having the classroom asking questions like: who is this player? When the record was recorded? Who else is playing on this record?   Experiences like that really played into the main reason we developed this app.  I was reminded of that moment, and I wanted to try and replicate something close to thumbing through a record collection, or the bins at a record store, with this app.  When we started talking about creating the Blue Note app, I immediately thought of what I loved about discovery in the past, and how I could structure the data in a similar way to replicate this process. There are a lot of improvements that can be made to the online discovery process, overall.  If you look for John Coltrane’s catalog online, for example, it’s a mess.  Nothing tells a story or takes you on a journey.  It’s like a jungle out there, totally convoluted.

MK: Can you talk more specifically about the process of building the app?

WG: Well, the idea really was in how to tell a story with finely tuned and structured data.  I started talking to developers, and found a company that we really liked in the U.K. called Retrofuzz.  They had done some apps that I was impressed with, and we decided to move ahead with them.  I had heard a lot of stories about how meticulous Spotify was from a development standpoint, and I wanted to use a developer that had been through the process already. It was important for me to move through the development process as smoothly as possible.  I went to Retrofuzz with a problem set covering 73 years of music, and I laid out all the data points I wanted to cover: artists, instrumentation, recording dates, and the ability to tag releases by simple genres.  It was important to me that we kept it all simple, as I didn’t want too many jumping off points, and I didn’t want to pigeon hole artists.

The other thing I asked Retrofuzz to do was to override some of the incorrect cover art in Spotify.  With old recordings, where records have been re-licensed, many times the album cover is not the right, or original, one.  I wanted the experience in the app to be as authentic as possible, so we created a way to override any incorrect album covers in Spotify, and replace them with the original vintage covers.

With this app, the whole thing really is a very specific look at the artist’s output on the Blue Note label.  Of course with the merger of EMI and Universal, it’s possible to include the Verve and Impulse catalogs into the app, but we wanted this to be just about Blue Note. We wanted this to be accurate for purists, which there are many of in this genre.  For example, Cannonball Adderley has only one record on Blue Note, but many on Capitol. Our label-focused approach gives us the ability to put a long tail look at this thing, too.  We can tell the Pacific Jazz and Capitol Jazz story at some point, for example. Michael Cuscuna was also a big inspiration.  Michael is one of the foremost experts on the Blue Note catalog, and we wanted the type of detail that he would think was appreciate.  We actually used his discography as a resource for plugging in a lot of the data.

 MK: As a discovery vehicle, the app is outrageous.  I’m a big fan of the label, and I’m surprised with how much I didn’t know about the catalog.  I’m constantly going down a different rabbit hole around some of the sidemen on these records, and being turned onto releases I had no idea existed. It must have been a serious process getting all this information into the app. What was that like?

WG: It was definitely a time consuming process to create that experience. When Retrofuzz finished building the CMS, I spent a full month plugging all the appropriate information in, all by hand.  Our team had to set up the full artist info, like the first and final recording dates, sideman info, and so on. During this process, I had my wife yelling at me for spending so much time on it, and I had to say, “Trust me, this will be really cool!”  I ended up bringing in some pf the rest of my EMI Digital Marketing team to input the data at one point. The problem there was that because I knew the info topically, I could blast through setting everything up quickly using Michael Cuscuna’s book to see correct sideman and instruments, but for the folks I brought in who were not particularly familiar with jazz and the catalog, it was a slow and painful process. Should this be a this groove or tradition release? Is that an alto or tenor sax on this record?  This was the sort of thing I knew intuitively, but for the folks I brought in, it was really tough. I actually used All Music and Wikipedia in the background as resources to double check everything, and in fact, all the bios are pulled from All Music.

MK: Is the entire Blue Note catalog available in the Spotify app?

WG: No, not all of the records are up on Spotify, and a lot of that is because of licensing issues.  For example, some records are not available from a licensing standpoint in U.S., and for us, we need to look at the ones that can be cleared quickly going forward.  There are a couple of inconsistencies that we have to work on, too.  Some records simply never became fully finalized in the US, even though we have a worldwide clearance.  It could have been a simple issue of the person setting up the licensing left the company, or forgot. There are some glaring omissions that we need to focus on, like Joe Lovano and John Scofield.

MK: Are you approaching this app as a potential revenue generating opportunity, or is this simply about awareness of the catalog?

WG: There is no one looking at this app saying “well, here’s comes the money!” right now, but I can say that this could change with the scale.  The analytics associated with engagement and usage far exceed our expectations.  We just launched the app, and right now we are getting around 16,000 visits a day, and the average use time is over three hours.  That’s amazing engagement.  I think that as this adoption grows, the app could drive revenue.  Right now it’s definitely an awareness tool, and a great opportunity to put the Blue Note brand and catalog back into the forefront. There are some really good things happening with the label right now.  Don Was is now involved as the President of the label, and he is driving A&R.  He came in last fall, and he brings great energy to what we are doing.

MK: Are you leveraging the work that Retrofuzz did for the CMS in other ways?

WG: We are.  Retrofuzz built a really amazing database,  and we are actually in the process of refining the Blue Note web site to be driven by this data base. The Spotify app will inform the Blue Note web site and we’re planning using the same CMS to power Bluenote.com.

MK: What’s your favorite part of the app?

WG: I’m a fan of the Blue Break Beats section. This is the area where we pull in content illustrating who sampled what in the Blue Note catalog.  When we were conceptualizing the app, I had an epiphany:  there has to be data out there that shows every sample from a Blue Note record, and where that sample occurs. There’s a great site out there called “Who Sampled,” and I ended up connecting with them and letting them know what I was looking for.  They came back with hundreds of Blue Note samples, and we integrated all of this into the data structure.  Everyone from Jay Z, to Madonna, to 90s hip-hop artists have sampled the catalog, and we structured the Blue Break Beats area to align the original recording and the sample side by side.  For me, it’s simply another great discovery tool, and I think the usage in the app is so high because there are so many ways to discover music within this app.  It really illustrated the possibilities of how music discovery can be in an online environment.

You know, Spotify loves the app, too.  They recognize that this steps the bar up, and they expect folks to improve upon this concept to offer more opportunities for deep music discovery.  For me, it’s nice to be part of a game changing development. I’ve done a lot in my professional career, but nothing has gotten the press or accolades that this app has gotten.  It’s a pretty neat thing to be involved with.

Grizzly Bear was blowing up my Twitter feed this weekend after asking the question: what’s up with Spotify’s payment model?  It’s not an uncommon question lately, likely due to the fact that no one can seem to pinpoint how exactly the service pays artists.  Grizzly Bear themselves claim to get about $.001 per stream, David Harrell from Digital Audio Insider averages closer to $.004 per stream over the past three years, and this infographic, which circulated widely a while back, indicates that artists on a label are paid $.00029 per stream.  So, what’s the deal?

The confusion is warranted – the interactive streaming payment model that Spotify, Rdio, MOG and Rhapsody use is less transparent than the permanent digital download model that iTunes employs, for example. The payments are variable, and payments are made to labels who distribute to their artists directly, which further obfuscates the process.  That being said, the subscription based interactive streaming model will likely continue to play a growing part in the future of music consumption.  As the most recent 2011 RIAA Year-End Shipment Statistics outline, subscription services were up 18.9% in volume from 2010, and up 13.5% in revenue. Small numbers compared with CD and permanent digital download (MP3) revenue and units shipped, but impressive when you consider that one of the major interactive streaming companies, Spotify, has only been active in the US since July of 2011. As we move towards a world where interactively streaming music will be one of the many growing options that consumers will choose to listen to music, it makes sense to understand how the financial process behind subscription interactive streaming works.

===

I’ve known D.A. Wallach for several years, after first interviewing him for my Online Music Marketing with Topspin course.  In addition to being the lead vocalist and songwriter in Chester French, D.A. works with Spotify as their “Artist in Residence.”   Below is a transcript of a conversation I had with D.A. about Spotify’s payment process.

Mike King: I feel like there’s a disconnect between artists and Spotify in regards to the mechanics behind Spotify’s payments.  There’s a lot of discussion about the deals that Spotify has with the major labels, and how the payments to indie artists vs. the payments to major labels is lopsided. Is there any difference between the payments Spotify makes to the majors, and the payments Spotify makes to indie labels or indie artist services like TuneCore or Cd Baby?

D.A. Wallach: We treat payments to indies and major labels the same way.  Let’s take a step back first and talk about some of the basics with the service.  We make money in two ways.  We make money through advertising to free users, who have access to Spotify only on computer.  The service is interrupted by ads, and the functionality is a lot like YouTube.  There is no mobile option for free ad-supported users, either.  Second, we generate revenue from selling subscriptions.  In the U.S., a subscription is $120 a year.  In the U.K. it is ₤120 a year, and in the E.U, it is €120 a year.

We aggregate all of this revenue from these two streams, and distribute back 70% in royalties based on a pro rata share in accordance with the popularity of a piece of music.  For example, if one of your songs has been streamed 1% of the total number of streams in a month, you will get 1% of the 70% of royalties we pay out to rights holders.  We pay this out to whomever owns the music.  If you are going through TuneCore, we’ll pay them directly, and because TuneCore takes no percentage on the revenue, whatever we pay TuneCore on behalf of the artist goes directly to the artist.  If you are signed to a label, we’ll pay the label, who is then responsible for paying the artist based on the contract the label has with the artist.

MK: Can you talk a little bit more about the revenue split between publishers and master rights holders?  How is the 70% of revenue you pay out split between publishing and the master side?

D.A.: With the publishing side, it’s a bit of a complicated formula. The rates are statutory, and have been negotiated with the PROs.  [NOTE: A good starting point to understanding how the interactive streaming services pay publishing royalties is this article from The Future of Music Coalition. Boiled down to basics, interactive streaming services pay a mechanical royalty rate of 10.5% on the revenue they generate, MINUS any amounts for performance royalties. In other words, services like Rhapsody and Spotify are subject to both a mechanical and performance royalty, but the entire compensation for songwriters and publishers from any limited download or interactive streaming site is "capped" at 10.5% of the site's revenue.]  In the U.S., we use Harry Fox as our service provider, and they do the distribution to the publishers.

MK: So there are no differences between what you pay a major label and what you pay an indie label?

D.A.: We have thousands of deals with all sorts of entities including distributors like the Orchard and TuneCore, the majors like UMG, and thousands of other independents.  The basic principle of the deals and the rough numbers are within a small margin in all of these deals. At the end of the day, the indie artist is not at a disadvantage compared to a major label artist, and we feel that all artists are being compensated fairly.

MK: Why do you think that there is so much confusion about how Spotify pays artists, and a general concern from artists about the payments they are seeing?

D.A.: I think there are three answers to this question.  First, we’re not a big company.  We have four million subscribers, and 15 million active users at the moment.  These are satisfying numbers but they are not staggering numbers.  We’ve paid out a good amount in royalties so far, close to $200 million dollars.  I think that people are comparing what we are doing to iTunes, which is not a legitimate comparison.  iTunes is orders of magnitude larger than we are.  People are expecting to see iTunes numbers, but we’re not there yet.  The second answer is that people need to transition from unit-based thinking to consumption-based thinking in terms of royalties.  We feel the metric of success should be based on how many people are listening to your music over a period of years, as opposed to looking at how many units are shipping in one week.  People are used to seeing big numbers from a unit-based model, but that’s really front loading what is happening. Comparing iTunes sales with Spotify payments over a two month period of time is not a great way to look at things.  What we are trying to create is a system in which you earn royalties forever for good music, and the time horizon is simply different than what folks are familiar with now. One can actually think about a download sale as a down payment on all future listening that a fan will do. If you took the effective per play rate that I’ve paid for every time I’ve listened to my Dark Side of The Moon CD, it would be trivial compared to what I’d have generated if I’d done all that listening on Spotify. The third answer is that it’s a confusing model since it is unfamiliar.  There is no fixed play rate, and as we grow, our royalty base constantly expands, driving higher and higher royalty payments. There is also confusion that arises from the fact that we pay royalties (just like iTunes, by the way) to whomever owns the music. In the case of a band on a label, the label generally mediates the accounting of those royalties. My band was on UMG, and when I look at my statement, as one example, it’s confusing. I personally hope that the conversations about Spotify royalties actually lead to efforts at increasing transparency in the entire digital music system. We’re very proud of the hundreds of millions in royalties that we’ve been able to pay out to the creative community, and we want the flows of revenue to be clear to artists.

MK: So to reiterate, as Spotify grows, the pool of revenue will increase, and the royalty rate will increase for rights holders.

D.A.: Yes, the larger Spotify gets, the larger the royalty rates should be. The royalty rates we’ve paid out have been growing at an exponential rate, and we expect this to continue.  If we can get to the scale of Netflix – which has 20 million subscribers – we estimate we’d be paying out to artists what iTunes is paying out on a year to year basis. This is a simple calculation based on the average download consumer spending $60 a year with iTunes, and the average premium subscriber paying $120 a year with Spotify.

MK: Do you think that Spotify is cannibalizing other revenue streams, such a downloads or physical sales?

D.A.: In no market where we exist has there been any data illustrating a downturn in physical or digital sales.  Many labels view us as an extra check, as a purely additive income stream, and I think this is an accurate way to think about what we are doing.  Our main demographic is 18-29 year olds, and in many ways, this is a generation who has never paid anything for music.  They grew up with P2P services, and most of these folks are paying for music for the first time in their lives.  It’s found money for artists and rights holders. On an individual artist level, we’re paying out royalties of $200-300 thousand dollars a month for some of the biggest acts.  The bottom line for us is that we have paid out nearly $200 million in royalties and we feel we are making a real contribution back to the music business. Not all artists are earning big checks, but this reflects a small user base and their relative level of popularity. It is also true that if not a lot of people buy your album on Amazon or iTunes, you won’t be seeing massive payments, either. That being said, we are a newcomer to the market, we’re making huge strides, and it will only get better.