Digital Aggregators and Spotify

You may have heard of “The Music Business Podcast.”  It appears to be produced by a fellow named Kyle Bylin.  His bio says “Kyle Bylin is a user researcher at SoundHound Inc.”  Soundhound is a Spotify partner and is well integrated into the Spotify app.  See the Spotify FAQ.  Soundhound also has a voice recognition app that seems to be a direct competitor with Apple’s Siri (Siri with a PhD).

Why do I mention this?  The Music Business Podcast (link to Soundcloud) has an interview with an artist named David McMillin entitled “It’s Time to Stop Hating Spotify” that has some flaws.  One thing Mr. McMillin definitely got right is that Spotify needs to do a much better job of converting free users to subscriptions and he did point out the trap that artists are in because Spotify was allowed to get away with offering the free tier in the first place.

However, Mr. McMillin and Mr. Bylin’s podcast proceeds on the premise that I think can be fairly summarized this way:  If you are an artist who has CD or download sales that are displaced by streaming, i.e., something to lose, you’re going to complain about your micropenny streaming royalty.  If you are also signed to a major label, the reason you’re complaining is not because of the micropennies but because your label is screwing you.

If you’re an independent artist who never had significant CD or download sales, i.e., nothing to lose, then you’ll be willing to accept your micropenny streaming royalty because it’s great promotion for….something.  And of course, don’t get the wrong idea, because Mr. McMillian would love to have a major label record deal.

If you stay with the interview, the typical ageism creeps in.  If you’re one of those artists with something to lose, you’re old.  If you don’t, you’re young.  That’s only a slight gloss.  Then it comes out that Mr. McMillin is 31 and his interviewer is 35.

These are obviously people who have never been to Silicon Valley.  A hot tip:  31 is OLD in the Valley.  Why do you think Peter Thiel, the techno utopian monopoly promoter, is paying STEM-high schoolers to develop apps for smartphones instead of going to college?

So here are some other flaws.  As you read the rest of this post, just be clear–I’m not saying that Mr. McMillin was paid by Spotify to do the Music Business Podcast interview.  I’m not saying that Mr. McMillin was paid by his distributor Tunecore, either, that he spoke well of in his original blog post.  I’m not even saying that Mr. Bylin was paid by Spotify to promote Mr. McMillin, or that if Mr. Bylin did the same thing as an employee of an FCC licensed radio station that he could possibly be charged with the crime of plugola.

I’m just saying that Mr. Bylin’s apparent employer is highly integrated with Spotify.  Whether that colors his editorial decisions when he promotes messages that are also promoted by Spotify without acknowledging the obvious is between him and his conscience.

(As an aside, Mr. McMillan and his interviewer showed a stunning lack of awareness about book piracy–obviously never looked for pdfs on torrent sites or fake medical text books.  So figure that out.  Also Mr. McMillin does not seem to be aware that Taylor Swift took down her catalog from YouTube and limited the availability of her current videos.  (PS not to “judge”, but like the reason Soundcloud doesn’t like pay artists is like because they are like ripping off like artists, however like useful the like site may like be.))

But let’s not be nitpicky.  First, it’s interesting that the interviewer said “people” are suggesting that labels are essentially at fault for low Spotify payouts.  Let’s be honest–the “people” who are making that “suggestion” start with Spotify itself.  Daniel Ek is screaming it from the rooftops.  The “people” start with the one that’s partnered with Mr. Bylin’s employer.  Those people.

Aside from focusing on major label artists without mentioning that those artists get advances that need to be recouped and aside from never discussing the mega valuation of Spotify, they only tangentially mention that Spotify’s all-in royalty rate is horrendously low.  (“All-in” meaning the total gross royalty, artist share and label share combined.) For example, Zoë Keating who disclosed her royalty statement also gets an extraordinarily low all-in royalty from Spotify.   Mr. McMillin also likely gets an all-in royalty rate as his band is paid through their aggregator, Tunecore–as he’s only too happy to tell you.  To the point that he sounds like he’s a spokesman and one could get the impression, albeit erroneously apparently, that there’s something in it for him.

A point that both miss is that the aggregate value of all music on streaming services is being transferred in large part to equity holders in the service which in part explains why Spotify can raise huge amounts of venture capital.  This is especially obvious with Spotify but is also true of YouTube (and Professor Jonathan Taplin might argue is true of Silicon Valley companies generally).

In a prescient 2008 book review of Nicholas Carr’s The Google Enigma (entitled “Google the Destroyer“), antitrust scholar Jim DeLong gives an elegant explanation:

Carr’s Google Enigma made a familiar business strategy point: companies that provide one component of a system love to commoditize the other components, the complements to their own products, because that leaves more of the value of the total stack available for the commoditizer….Carr noted that Google is unusual because of the large number of products and services that can be complements to the search function, including basic production of content and its distribution, along with anything else that can be used to gather eyeballs for advertising. Google’s incentives to reduce the costs of complements so as to harvest more eyeballs to view advertising are immense….This point is indeed true, and so is an additional pointIn most circumstances, the commoditizer’s goal is restrained by knowledge that enough money must be left in the system to support the creation of the complements….

Google is in a different position. Its major complements already exist, and it need not worry in the short term about continuing the flow. For content, we have decades of music and movies that can be digitized and then distributed, with advertising attached. A wealth of other works await digitizing – books, maps, visual arts, and so on. If these run out, Google and other Internet companies have hit on the concept of user-generated content and social networks, in which the users are sold to each other, with yet more advertising attached.

So, on the whole, Google can continue to do well even if leaves providers of is complements gasping like fish on a beach.

You could replace “Google” with “Spotify” in this passage and I think it would work as well.  (And, of course, Google recently joined Spotify’s board of directors, knowing a good commoditizer when they see one).  These are facts that seem to escape Spotify boosters.

For whatever reason, the fact that “people” are getting extraordinarily more wealthy than any artist on their service or even all artists on their service (Daniel Ek $250 million net worth, Tim Westergren making $1 million a month or more) doesn’t seem to bother Mr. McMillin (or Mr. “McMilion” as he’s ironically referred to by the podcast–spellcheck, people).

But here’s the main flaw in Mr. McMillin’s reality: He seems to think that the reason he does better than a major label artist is attributable to his distributor who he just pays on an annual flat fee basis and not that he gets an all-in royalty from Spotify or other streaming services.  The idea that an annual flat fee distribution fee is somehow not a distribution fee is really hard to accept.

A flat fee distribution fee that the artist pays rain or shine is simply another way for the distributor to push 100% of the distributor’s risk onto the artist.  If the distributor takes a distribution fee that is dependent on sales, the distributor only makes money if the artist makes money.  This seems to be pretty obvious.  And frankly a far better allocation of distribution risk.

Why?  Because the artists who do best on a purely distribution fee basis are those who sell a lot.  As you will see below, those artists who don’t sell a lot are essentially paying a colossal distribution fee and are also protecting the distributor’s downside risk.  Which can also translate into the distributor not being incentivized to fight for the best terms in their own deal with, say for example, Spotify.  I’m not saying they did or didn’t just asking the question: How would you ever know?

Why an artist would be proud of the fact that they have taken 100% of the distributor’s risk onto themselves is beyond me.

Those same flat fee distributors also don’t allow you to audit and won’t show you their direct deals with any digital service, so no one knows whether they pass through the terms or if they are keeping a vig for themselves off the top.  That’s not an accusation, it just seems to be the fact–no one knows.  If that’s wrong, please send me the terms and we will post them.

Here’s an excerpt from a post about digital distribution deals I wrote a few years ago that addresses this issue (Digital Aggregator Deals: Is the New Boss Worse than the Old Boss?).  I think it’s still relevant with one caveat:  I used a download model for comparison purposes.  If you use streaming revenue instead, make sure you have a scientific calculator.  You’ll need more places to the right of the decimal.

Commission Rates: Percentage vs. Subscription

How the aggregator is compensated is also an issue of concern. In the traditional model, the aggregator took a percentage of sales as their compensation. This meant that the aggregator only made money if the artist made money. Some aggregators charge a flat fee on some basis (such as a per-retailer basis) instead of a percentage, or an annual flat fee.  This makes the aggregator deal more like a subscription model where your credit card is banged every year for a magazine subscription.

Each model has its strong and weak points. The percentage model pays the aggregator a percentage of your gross revenue that the aggregator collects regardless of whether they are making an effort to stimulate sales (which few of them do in any event regardless of how they are compensated). However, under the percentage model the aggregator only makes money if you make money, so at least the incentives are aligned. The percentage should be low (15% or so is fairly typical) to take into account that the aggregator  has lower incremental costs over time of maintaining content in their catalog.

The flat fee model has the artist pay the aggregator a fee for distribution instead of paying the distributor a percentage. While this is attractive from the point of view that the artist knows what their distribution costs will be up front, it also transfers all of the risk of distribution to the artist. In order to determine which is the better model, the artist should compare their most favorable percentage based offer to the flat fee model and see what the breakeven point will be. Try using a formula like this and solve for “X”:

[Flat Fee]/[percentage] = Gross Income

Gross Income/wholesale price = breakeven units

or, for example if the flat fee is $100 and the comparable distribution fee deal is 10% (which would be very low but this is an illustration):

$100/.10 = $1,000 (Gross Income)

$1,000/$0.70 = 1428 units (rounded down)

In the example, a $100 flat distribution fee is equivalent to a 10% distribution fee model if you sell 1,428 units at a wholesale price of $0.70 (a typical wholesale price per track for permanent downloads). That means that if you sell exactly 1,428 units you will be indifferent between the two models. It also means if you sell fewer than 1,428 units, you will be better off under the percentage model. If you sell more than 1,428 units you will be better off under the flat fee model.  (You could argue that the units would be 10% higher to get to a net number to the artist, but we are trying to keep it simple.  That difference would be another 159 units [(1428/.90)-1428], rounded.)

This example is only for one accounting period and only uses one revenue stream–permanent downloads.  It is more likely that you will see blended revenue streams, but we factor out limited downloads and streaming because permanent downloads are the overwhelmingly dominant revenue stream for most artists.  That may change over time or be different for you.  Also, as you extend the distribution costs and revenues over longer periods of time (with additional flat fee payments per year under the subscription distribution model), your results may vary.

To take another example of the flat fee model, what would the flat fee equate to under the percentage model at 500 units at a wholesale price of $0.70 (a typical wholesale price for permanent downloads)?

[Flat Fee]/[Gross Income] = Distribution Fee as a percentage

$100/[(500) x ($0.70)] = 28.6%

Under these assumptions, a 28.6% distribution fee for an accounting period would be in the astronomical zone for a digital aggregator–who should be getting around 15%.  It would even be on the high side for a major label distributor who was also giving signficant (and expensive) marketing, PR, sales and radio promotion support.

But these are just assumptions to illustrate the issues.  In any of these examples, you will need to use your own projections on sales, wholesale price and configurations in order to get a projection that is personalized for you.