Do Freemium music services offer greater revenue per user than the freemium combo of Radio and Sales?

The purpose of this story is rather simple:

First, let’s treat the music industry like a “freemium” music service with both paid and free components. And, using the data we gather, estimate the average revenue per user for this freemium experience.

In fact, the music industry has always been a freemium music experience at the macro level: Free, ad-supported services like radio existed to both (a) make some money selling ads and (b) hopefully make even more money selling records. 

Then, let’s find an alternative freemium experience—the so-called counterfactual—that also attempts to (a) make some money selling ads alongside a free music experience and (b) hopefully make even more money selling a paid version of that experience.

I don’t know about you, but I think a service like Spotify might suit our needs for the thought experiment at hand.

Finally, since the RIAA and Spotify have been kind enough to release figures from 2015, and both ASCAP and BMI have released figures as well (as far as SESAC, we can make a guess), we are in position to compare the average revenue per user for the US Music Industry  (one freemium economy) against that of Spotify (another freemium economy).

Why would we make this comparison? Because as far as I can tell, a great proportion of the FUD related to music services reduces to a desire to just have the world as we used to know it — a simpler time, with free  Radio and paid CD or Download Sales, and none of these pesky music services.

So, let’s compare these two worlds using a similar metric and the data at hand to see which workd we should prefer.

What we are going to find is not necessarily what we might expect:

A Freemium music service like a Spotify — one wherein free experiences hopefully convert to paid subscribers — may actually offer a greater average revenue per user than the “freemium” industry we knew so well — one wherein free radio spins hopefully convert to paid sales.

$25.08 = Average Revenue Per User, US Music Indsustry

$26.11 Average Revenue Per User, Spotify.

What follows are the numbers on the back of the napkins through which the above results were estimated. Please feel free to digest, debate, and even debunk.

US Music Industry

On the back of this napkin, we have the 2015 results reported by the RIAA, ASCAP, BMI (a mix of 2014/15), and a decent guess at SESAC.

Total Active Users (TAU) in this world, just like active users in the online world, would be anyone who connects with our “service” — free or paid — during a thirty day period. Thus, a simple way to impute Active Users in this “real world” is to just use the number of listeners who connect to radio each month.

And remember, a simple way to estimate Average Revenue Per User (ARPU) is to simply divide Total Revenue from a service across all version of that service (free and paid)  by the Total number of active users.

$7,015,900,000 (Total digital and physical, paid and free revenue, RIAA)

(-) $1,218,900,000 (Revenue from paid music services — removing our other economy from the mix)

(+) $415,000,000 (estimated Total royalties paid by Radio, across ASCAP, BMI, and SESAC)

(=) $6,212,000,000 in Total Revenue

(/) 247,706,000 (Total audience, average, according to RAB/Nielsen)

(=) $25.08
Average Revenue Per Music User, for the US Music Industry.

Spotify

Estimating Spotify’s ARPU is a bit easier since, well, the service a closed economy with a single source for information — the company’s supposed financial filings, recently revealed (thank you, Mr. Ingham, and Music Business Worldwide).

$1,945,322,000 (Total Spotify paid and free service revenue)

(/) 74,500,000 (Total active users, average across 2015)

(=) $26.11 
Average Revenue Per User,for Spotify.

WHAT?!?!

I have to admit, the first time I ran these numbers, I was boggled as well. But remember, Radio’s conversion rate (the ratio of spins to people to sales) is really, really low; So low, that the free-to-paid conversion rate for services like Spotify—however disappointing to some stakeholders—is still orders of magnitude greater than that from Radio.

BUT…

But wait — those streaming revenues you pulled out from industry revenue could have been sales, so put them back in!

OK… But where that line of reasoning will take you if you really follow through is quite simple: A wayback machine to the numbers before 2003 — the days before the “new” music experience that are feared to cannibalize sales (e.g., webcasting, services, even the sales of downloadable singles). And a trip on that wayback machine is for another day.

But wait — those Spotify figures are global, and include both small and large countries, at a variety of prices, some of which are not comparable to prices in the US Music Industry.

That’s right, they sure do and they sure aren’t. But, we would expect this bundle of countries to be worth less than the US. And yet, that is not what we see. And a comparison using a bundle of date from a variety of countries is for another day.

 

Flood Stage Economics: The subtle and often unspoken dangers in the displacement debate

Recent work by Dr. Aguiar and Dr. Waldfogel has made its way from a working paper to a technical report released by the Institute for Prospective Technological Studies, one of the Joint Research Centres within the European Commission. The work, “Streaming Reaches Flood Stage: Does Spotify Stimulate or Depress Music Sales?,” (link to PDF) attempts to determine whether any displacement (i.e., cannibalization) of sales occurs as a result of increased adoption of streaming services, services which in this case are considered to be wholly represented by Spotify.

Most news outlets appear to be simply reporting the findings as selectively summarized in the abstract of Aguiar & Waldfogel’s paper. This selectivity amidst the findings is unfortunate (or perhaps, welcomed) as the abstract only reports a portion of the findings from the research.

Furthermore, there is an issue with the research methods that I reckon we need to consider given the so-called called “Displacement Debate” (previously, the Cannibalization Conundrum) continues to heat up.

The Short Version:

Upon closer examination its clear that the findings from the research are far more complicated than initially apparent. In fact, certain portions of the work find significant and negative (in terms of financial consequences) displacement of Track Sales as a result of Spotify adoption. While another portion of the work (that portion most alluded to in the abstract) suggests a more neutral form of displacement. And yet, why one set of findings would be more reliable than the other is not entirely clear.

Furthermore, there is a subtle issue embedded in the method that led to the more neutral findings—an issue that may understate the displacement reported in the abstract. The authors use a sample of the Top Streaming Tracks to derive conclusions in the Total Digital Tracks Sales market, but don’t take into account the fact that the Top Streaming Tracks are a much smaller proportion of Total Track Streams than the Top Sold Tracks are as a proportion of Total Track Sales.

And while any hint of more substantive and negative displacement might seem to provide the smoking gun among displacement theories, there would seem to be good reason to question any of this smoke.

Ultimately, the factor that it seems we really need to think about is that the work technically asserts—via the data employed—that Spotify is the only force behind both (a) displacement resulting from streaming services as well as (b) any earned revenue from these streaming services that might make up for any observed displacement. Such an assertion is especially challenging in the US, where Spotify—while significant in stature—is not the only significant source of streaming demand or revenue.

In other words, we don’t know which portion of any lost revenue in the sales market could be attributed to Spotify, exclusively. And yet, the Flood Gates paper essentially treats the situation as if the entirety of any decrease in unit sales can be explained by adoption of Spotify alone.

As a result of this final limitation, drawing dramatic conclusions from these Flood Gate findings may ignore the subtle and often unspoken dangers in the displacement debate.

The Long Version:

Simply stated as simply as possible, the findings from this research that are most directly referenced in the abstract are a function of streaming data gathered from 86 tracks found within the top 200 tracks streamed on Spotify, within the US, over a period of time April 2013-March 2015. These streaming data are compared to changes in (I believe) total digital tracks sales in the US, over the same period. The results suggest that any lost sale is associate with 137 additional track streams.

While its not entirely clear at this point, the conclusion of the authors is that any displacement is revenue neutral. “Given our preferred estimate of the rate of sales displacement (-0.0072) and a payout per stream of ps =$0.007, the growth in Spotify appears to be essentially revenue-neutral for rights holders.

That said, if we multiply 137 by what is supposedly the per-stream average from Spotify ($0.0072), we get $0.99—which suggests we make 99¢ in streaming royalties for every 82.2¢ (average Sales royalty) we might lose in Track sale—a gain of 17¢. Ka-ching!

However, another set of findings within the paper lead to a very different conclusion. When streaming data from approximately 714 Top Tracks from 2013 that could be matched by Song and Artist across both Streams and Tracks Sales, the authors find evidence of noticeable displacement.

In the case of these data, the authors found that every 43 additional streams on Spotify seemed to be associated with a displaced track sale. In this case, we make 31¢ in streaming royalties but forego 82.2¢ in track sales royalties—for a total loss of 51.2¢. Uh-oh.

And so, and as stated above, the findings from this research are a bit more complicated—if not conflicted—than appears to be widely reported.

The above said, there is a significant difference in the approach to these two findings: The former approach has a nice matched set of Tracks in the Streams and Sales domain from which to draw conclusions. The same Songs/Artists are compared across the Streaming and Sales domains. Using this matched set, we find a scary sort of displacement, just in time for Halloween.

The latter approach involves a “sample” of the Top Tracks in the Streaming market and draws conclusions, using this special group, about the link between Streams and Sales, in general. Using this not-so-matched set, we find a more benign sort of displacement.

[scratch head here]

Aguiar & Waldfogel work to account for any difference by essentially multiplying/dividing results by the ratio of the Total Streaming market account for by some group of Top Track. I don’t know if this single adjustment is enough.

What concerns someone like me is the extent to which we can draw conclusions using what feels like a “sample” that is comprised of Top Tracks in the Streaming market, without taking into account the extent to which the Top Tracks in the Streaming market are far less influential than the Top Tracks in the Sales market, when it comes to the Total Tracks Sold or Streamed (in their respective markets).

In other words, the roles these Top Tracks play in the Streaming versus Sales markets, in terms of the total distribution of sales, is quite different. For example, using Nielsen top-level numbers as reported in their 2014 report (link to PDF)…

  1. The Top Ten CD albums account for 6.624% of total CD Album Sales.
  2. The Top Ten Digital Tracks account for 3.759% of total Track Sales.
  3. The Top Ten On-Demand (Stream) Tracks account for 0.98% of total Track Streams.

And so, there may be reason to account for the fact that any findings in the not-so-matched set used for the Flood Gates work may need to be adjusted for the fact that any displacement of these similar Top Tracks would be more influential in the Sales market than in the Streams market. A simple way to see the consequences of this difference in ratios might be:

Assume all Streams of the Top Ten Tracks were somehow displaced (i.e., erased) by an unexplained force. The erasure of  these Top Tracks would only reduce total Streaming royalties by 0.98%
Now, assume all Sales of the Top Ten Tracks were similarly displaced by an unexplained force. The erasure these Top Tracks would reduce total Sales royalties by 3.759%
Top Ten Tracks are nearly four times as important to, or influential in, the the Sales market as compared to the Streaming market. It seems we may need to think about this “sampling” issue, particularly given we can use the back of a napkin to bring the findings of the not-so-matched set very close to those of the matched set, simply by taking into account greater influence in or consequences of displacement in the Track Sales domain.

Ultimately, and as stated in the introduction, it seems we really need to think deeply about the fact that this research technically asserts—by way of the data selected—that Spotify is the only source of both (a) displacement resulting from streaming services as well as (b) any earned revenue from these streaming services that might make up for any observed displacement. Such an assertion is especially challenging in the US, where Spotify—while significant in stature—is not the only significant source of streaming demand or revenue.

Frankly, we don’t know which portion of any lost revenue in the sales market could be attributed to Spotify, exclusively. And yet, the Flood Gates paper essentially treats the situation as if the entirety of any decrease in unit sales, as well as offsetting royalty revenue can be explained by changes in the adoption of Spotify alone.

As a result of this final limitation, drawing dramatic conclusions from these Flood Gate findings may ignore the subtle and often unspoken dangers in the displacement debate.

Popularity versus Demand: Looking further into the fair distribution of streaming wealth

This post will be the last in a series of three conversations on the ways in which artist popularity, listener demand, and time spent listening impact any redistribution of wealth that could occur when comparing Total Pool and Per User (aka, User Centric) methods for royalty payouts from streaming music services. These are the models that folks like Laguna, Lowery, Mulligan, and others are now debating if not taking a position.

In the end, any resolution of this debate over the streaming models will likely have more to do with opinions and the influence of those opinions than it has to do with the numbers. Ultimately, I believe that this streaming payout concern boils down to two simple questions:

  1. Should the dollars I spend on a music service go into a big pool of money, from which the artists we all listen to are paid—based upon everyone’s listening behavior? or
  2. Should the dollars I spend on a music service go only into my own pool of money from which the artists I listen to are paid—based upon only my listening behavior?

The answer to the “Should” has nothing, ultimately, to do with who gets how much compared to whom. Instead, these are questions of process not payout. The nature of the consumer connection between dollars spent and royalties paid matters, regardless of whether artist payouts ultimately differ between these two models.

Frankly speaking, #1 looks like the way in which royalties from from venues such as Radio and TV. #2 looks like the way in which royalties flows from a sale—unique to each transaction.

Beyond these questions, in this post I am going to introduce a few additional scenarios that will hopefully surface the underlying lever(s) that ultimately determine whether any shift in payout model—regardless of opinions—will have any substantive (i.e., meaningful) impact upon royalty payouts.

What we are going to find is that these these two models—Total Pool versus Per User—result in substantively different artist payouts only when a systematic pattern exists. That systematic pattern is simple:

IF the most active listeners on the service listen to a different portfolio of artists from that portfolio to which the least active listeners on the service listen. The more unique the portfolio of artists listened to by these two user populations—heavy listeners versus lite listeners—the more dramatic the difference in royalty payouts as we shift models.

The above conclusion means that so-called independent or newbie or middle-class artists will only benefit from a shift in these two models if these artists happen to also be the primary interest of those listeners who spend the least amount of time listening to music. If, however, listener behavior is rather well-distributed across any and all “classes” of artists, a shift between these two model will have little impact upon artist payouts.

Furthermore, this conclusion operates in the context that free users on services such as Spotify stream a far lesser number of tracks each month than paid users—at least according to various new stories. And, these free streams mostly likely carry their own, per-stream payout (at least for artists signed to major rightsholders).

And so, if the least active listeners are actually just streaming the hits, while the most active listeners are streaming across the wide range of “classes,” some artists may find the redistribution of wealth under the Per User model goes counter to their expectations.

Let’s dig in…

Scenarios

In this section, I am going to introduce only four scenarios, simply because it seems these four scenarios are all it really takes to “tease out” this interaction between time spent listening and popularity.

In scenario one, all subscribers will listen to the same amount of music (900 listens), while there will be variety in terms of the portfolio of artists to which users listen—some users will listen to a wider range of artists than others. In scenario two, all artists will enjoy that same amount of demand (900 plays), while there will be variety in terms of the portfolio of artists to which users listen—some users will listen to a wider range of artists than others.

In scenario three, we are going to return to the basics—every subscriber will listen to the same number of tracks of each artists and the same number of tracks over all. And scenario four, we are simply going to twist scenario three to introduce the scenario that ultimately drives the difference in payouts between the two models: some artists less popular (i.e., in smaller number of playlists) and some listeners less active (i.e., listening to few tracks).

Note that the average daily listening hours (111), as well as the average payout per stream ($0.0078) in this example line up with the average released by Spotify through various new sources.

Scenario One

Varied Popularity, Equal User Demand, Identical Listening Hours

In this scenario, all subscribers spend the same amount of time listening to music, and in turn, stream the same number of tracks. Each artist, however, experiences very different levels of popularity (i.e., number of people listening) and demand (i.e., number of tracks streamed). In fact, Artist A experiences 29-times the number of streams as compared to Artist K.

 

VariedPop-Popu

As we might expect, Artist A enjoys the greatest proportion of the royalty payouts: 29.28%.

What we might not expect, however is that the payouts across all of the artists would not change if we shifted from Total Pool to a Per User method.

DifMethod

 

Essentially, even though some artists are far more popular than others (i.e., in a greater number of subscriber playlists) and far more in demand (i.e., streamed a greater number of times), because of the similarity in the time users spent listening to music, there is no different between the Total Pool and Per User approaches to payouts.

Scenario Two:

Varied Popularity, Equal Overall Demand, Varied Listening Hours

In this second scenario each Artists, A through K, experience that same level of demand across all Users, Q through Z. In other words, once all listens across all Users have been tallied, all Artists enjoyed the same number of listens: 900. While some artists are more popular than others (i.e. found in the greater number of listener playlists), they are all equally demanded, in terms of the total number of streams.

That said, User Z was the most active user, streaming the greatest number of tracks across the widest range of music, while User Q was the least active user, listening to the smallest number of tracks across only a single artist.

In other words, we took the variety in demand from the prior scenario and moved it to time spent listening in this scenario—to test whether Demand for Artists or Time Spent Listening ultimately drive any difference in payouts.

Varied-Distinct-Equal

Under a Total Pool approach to the royalty payouts, each Artist receives the same payout—$7.00—as each artist was equally in demand as the next. Under a Per User approach to royalty payouts, however, each Artist recevies a very different payout—ranging from $2.39 to $17.51—given total listener hours varied across all of the Users.

Equally worth noting, under a Per User approach some streams become worth more than others. Streams for Subscriber Q are worth 7.8 cents per stream, while Subscriber Z is paying out 0.27 cents per stream. Meaning a stream for Subscriber Q is, effectively, 28 times greater in value than a stream for Subscriber Z.

Is the music of Artist A worth more simple because it is preferred by the subscribers who stream less music? Are those moments with music truly more valuable to Subscriber Q? Or was he/she simply too busy to listen to music?

DifMethod - Listening

Artists E through K, popular among only those Users who listen to over an hour of music a day, all receive less money under the Per User approach. In particular, Artist K, uniquely listened to by the most active User, takes the most significant hit in a shift from the Total Pool to the Per user method.

Artist A, broadly popular among all Users, albeit not-so-in-demand by any User, experiences the greatest benefit in a shift from the Total Pool to a Per User approach. In the case of the Per User approach, it pays more to be more popular among those Users least passionate about listening to music on the serivce.

Note as well that Artist K experiences an effective royalty per play of $0.0027, about a quarter of a penny, when average royalty per play is actually $0.0078.

Essentially, Artist K receives a significant lower payout for being in very high demand from the most active set of users. Fair?

Scenario Three:

Equal Popularity, Equal Demand, Equal Listening Hours

Its time to go back to the basics and, in fact, back to where we started. In scenario three, every subscriber listens to the same number of tracks of each artists and the same number of tracks overall. Essentially, all variables are equal in this situation—popularity, demand, and time.

5

As we might expect, when all is held equal, there is no difference in payouts between the two methods:

6Call that a baseline finding.

Scenario Four:

Varied Popularity, Equal Demand, Varied Listening Hours

In our last scenario we are varying popularity and demand, while leaving equal the number of tracks any subscriber streams from any artist. In fact, not only are some artists more popular than others (i.e., found within a greater number of subscriber playlists), but also popularity is exclusive — some artists are listened to by some subscribers (thus the “zero”s in some artist/subscriber cells .

Perhaps most important for our purposes here, in this scenario the least popular artists happen to be found in the playlists of those subscribers who also happen to listen to the least amount of music.

7

What do we find?

8

Perhaps as no surprise by now, those artists who were not just more popular, but also in the playlists of the more actives subscribers, receive smaller payouts under the Per User method. However, those less popular artists, who also happened to be found in the playlists of the less (if not least) active subscribers, benefit most from the shift to a Per User approach.

Conclusion

For anyone who has actually bothered to read this far, I will just say the following.

Basing this decision over the payout method upon who wins or loses may be far less appropriate, and less loaded, then simply basing this decision upon which payout process seems the more direct for fans.

In either model, artists are going to see per stream payouts measured in the fraction of a penny. In fact, they are going to see variety in payouts—some streams become more more than others. A situation which will lead to a debate over whether and why the listens of less active users are worth more than the streams of more active users.

All the best, in the fairness debate.

 

 

Everything you need to know about why the Apple Brand may not be enough to move the needle for the Music Industry

Tomorrow, “it” happens. Apple’s long-awaited and much-hyped music service, Apple Music, will launch.

Along the way, we have had to endure near continuous use of the phrase “Move the Needle” from a music empresario (the same phrase used to describe a headphone company as well as a record label), negotiations via social media between a music empresaria and Apple, alongside an endless stream of “Apple Music, everything you need to know” articles (e.g., The Guardian, USA Today, Venture Beat, Wired, CNet, Mashable, The Week, Fortune, just to name a few).

Amidst all of the hype and hyperbole, there is some reason to wonder whether there may be three key consequences of the launch of Apple Music:

(1) A really big number of free trials. We could very well see numbers that map onto Spotify’s growth figures, as in tens of millions of free trials

(2) A measurable amount of churn among music services—or, at least among iOS devices. The question is whether the total market expands more dramatically than this inner churn.

(3) Some legal intrigue over the artist-generated and user-generated content on Connect — the sort of intrigue that will be a new experience for Apple.

The history of Apple Computer is, essentially, a 30+ year effort to perfect the art of hyperbole, with Job’s introduction of the Macintosh in January of 1984 setting expectations that remain through the decades. However, this time around there may be a few more gaps in the hyperbole, so we might as well as well address these gaps. And, we might as well just start out by stating the most obvious:

There is nothing truly “new” about Apple Music.

There, we said it. Nothing about Apple Music is truly “new.” Paid music services have been around since 2001/2, and since that beginning the gist and experience of music services has largely remained the same.

What has changed since 2001 is the presences and proliferation of smartphones, such as the iPhone, and the growth of the music service market, in part, tracks adoption of these mobile devices.

In 2015, however, very little distinguishes the proposed Apple Music service from those services by Rhapsody (the eldest), Rdio, Deezer, Spotify, Google Music, Tidal, and others. A great portion of this lack of variety is due to the context (and constraints) within which these services are licensed. Regardless, all services charge $9.99/month for the premium tier, all have (or will soon have) free trial periods, most can import playlists, many can match/store past music collections, all combined personalized radio with on-demand playback… we could continue.

The most striking feature that differentiates music services is whether the application has a white or a black background.

Will 800 million iTunes accounts spend more after spending less on music?

While Apple and other sources point to the market force that could be Apple’s 800,000,000 iTunes accounts, let’s just be frank: these accounts were already spending less music.

2014’s 8% decline in download sales, particularly singles sales, combined with a similar 8.1% fall in CD/physical sales, may be the cause of the music industry increased hope for (amidst still lurking fear of) music services.

However, the question of whether consumers see music services as a value proposition sufficiently compelling to treble (i.e., triple) household spending on recorded music is a big question, and has yet to be clearly answered.

Is family plan pricing enough of a “twist?”

If we were to be completely honest, the only twist unique to Apple Music, at least for the moment, is a $14.99 “family plan” tier that applies to up to six individuals—a twist that arrives by way of the Beats Music pricing via AT&T.

Across the other music services, such as Rdio and Spotify, additional family members are priced at $5 each, up to five accounts. In these schemes, a three-account family would cost $19.99 while a five-account family of five would cost $29.99.

There is some evidence out there that pricing matters when it comes to consumer adoption of music services (ahem). But, the music industry largely prefers to argue that the role of the service is to “sell” the fact that this on-demand access to music is a premium opportunity.

In reality, music service prices are falling. These prices drops simply occur through synthetic means — free trials and family plans.

Its unclear whether this family plan pricing will be seen as sufficient to shift music services from a premium niche to a mass-market opportunity. However, this pricing does present digerati families with a more compelling value proposition.

Can Free Trials convert Freemium Zombies?

While the industry prefers to pretend otherwise, the growth in paid music subscribers (particularly in the US, since 2011) lines up with the emergence of “freemium” offerings. However, the Apple Brand and a three-month trial may not be enough to suddenly convert the zombie accounts still walking the streets—those freemium accounts that went largely inactive after 90 days.

That last paragraph contains some controversy, so let’s use some data to justify the claim…

In the US of A, at least according to the RIAA, there were 1.3 million paid music service subscribers in 2005, 1.8 million 2007, 1.2 million in 2009, and 1.8 million in 2011. Suddenly, in 2012, there were 3.4 million subscribers and by 2014, 7.7 million subscribers.

Smartphone adoption was rising smoothly from 2007-2011, and the uptick in that adoption from 2011 to 2012 was really no different. Prior to 2011, most services had already dropped prices from $14.99 to $9.99.

What did happen in 2011 in the US? Freemium. Although, I guess you could say there is some evidence that simply being Swedish makes you more likely to subscribe to a music service.

Bigger market, or just more Churn?

In the near term, unless the Apple Brand has enough force to grow the market overall, all of the hype around Apple Music combined with the three-month free trial may lead to three things:

(1) A really big number of free trials. We could very well see numbers that map onto Spotify’s growth figures, as in tens of millions of free trials.

(2) A measurable amount of churn among music services—or, at least among iOS devices. The question is whether the total market expands more dramatically than this inner churn.

(3) Some legal intrigue over the artist-generated and user-generated content on Connect — the sort of intrigue that will be a new experience for Apple.

…More later.

Full disclosure: In honor of the launch of Apple Music I purchased a Nexus 6.

$0.0000955: The value of a Spin, per Listener, to Songwriters on US Radio in 2014

Now that ASCAP and BMI have revealed their revenues for 2014, and we have some data on SESAC (thanks to Moody’s), we can estimate the value of a Spin, per listener, to Songwriters from performances on US Radio, in 2014.

The estimates that follow are based upon figures that have been publicly released by a variety of sources (e.g., ASCAP, BMI, RAB, Arbitron, Nielsen) as well as my own estimates within these sources. And so, to the extent that these sources and my adjustments are correct, this estimate should be as close as plausible.

Importantly, I have shown my work, so please feel free to come to your own conclusion.

The Short Story

Given 2014 reported revenues, combined with audience and listening statistics, it would seem that the per performance, per listener value of the Spin on US Radio in 2014 would hover somewhere around:

$0.0000955… for the songwriter/composers/lyricists
$0.0000955… for the publisher(s)
$0.0000281… for the PRO (average across ASCAP/BMI/SESAC)
or
$0.000219 <- in Total for Songwriters, Publishers, and PRO

Importantly, since performing artists, musicians, and labels are not paid for performance on Radio in the US, the figure above describes only those payments made to or received by songwriters and composers.

Small numbers can be tough, but the above means that Songwriters receive about 1/10 of 1/10 of 1/10 of a penny for each performance, per listener, on US Radio.

Or, approximately $95.50 for a performance to 1,000,000 listeners.

The Long Story

$177,432.445 <- Estimate of ASCAP collections from US Radio*
$171,854,300 <- Estimate of BMI collections from US Radio**
$32,013,800   <- Estimate of SESAC collections from US Radio***

(A) $381,300,545 <- Estimate of Total PRO collections from US Radio

243,451,000 <- number of US Radio “active listeners,” weekly (Arbitron/RAB)

13.40 <- Number of listener hours (average) per week (RADAR/RAB, Nielsen)
52 <- Number of weeks per year
10.25 <- Number of songs per hour****

(B) 1,738,775,732,200 <- Estimate of Total number of unique performances

$0.000219 <- Value per Spin, per Listener for each performance (B /A)

Since the PRO takes some percentage of this value as Administrative fees, and these fees average around 12.85% between ASCAP and BMI, then about $0.000028 per performance, per listener, goes toward Administrative Fees paid to PROs (i.e., ASCAP, BMI, SESAC).

Publishers, afforded 50% of the payments, post fees, would earn $0.0000955.

While Songwriters (i.e., composers and lyricists) would earn the remaining 50%, or approximately $0.0000955 per performance, per listener.

Importantly, this oranges-to-oranges comparison is something not even the PROs themselves prefer to do. For example, BMI states that 98% of its monitored performances come from the internet, while 2% come from Radio—a discrepancy that results from the organization considering a performance to thousands,or millions, of people on the Radio (apples) as equivalent to a performance to a single person, or account, on the internet (oranges).

NOTES:

*ASCAP domestic revenue in 2014 should be around $655,969 million, a 6.7% increase (as reported) over 2013, when domestic revenue was $614.779 million. In 2013, Radio revenues were 27% of domestic receipts. $177,432,000 is simply 27% of $655,969,000.

**BMI is less transparent than ASCAP. When it comes to numbers, BMI releases only top-line revenue/distribution figures, and seems most excited about percentage increases in their Twitter, Instagram, and YouTube followers. And so, this estimate is a function of simply multiplying BMI’s top-line revenue ($977 million) by that proportion of ASCAP’s top-line revenue that can be attributed to Radio (17.59%).

***SESAC, being a private, for-profit company, is the least transparent of the US PROs. And so, this estimate simply takes a percentage of of total collections attributable to Radio, similar to that experienced by ASCAP, and multiplies it by Moody’s estimate of SESAC’s total collections in 2014 ($182,000,000).

**** Estimating the number of songs played per hour on US Radio appears to be a mix of Art and Science. 10.5 assumes about 40 minutes of music per hour, with songs that are just under 4 minutes in length.

Redistributing the streaming wealth: Understanding the links between popularity, demand, and time

The purpose of this post (which will be released as a Grey Paper) is to further inform the ongoing discussion over how best to distribute the royalties from subscription-bases music services. Over the next few pages we are simply going to think and type out loud, in an effort to understand how popularity, demand, and time (i.e., listener hours) interact to determine whether and how different payouts methods—in this case, “total pool” versus “per user” methods for distributing royalties—might lead to different outcomes for different populations of artists.

Rather than advocate that either approach to royalty payouts is somehow fairer than the other we are simply going to take a calculating look at a series of scenarios that present intentional variations in artist popularity, demand, and user listening hours—in the same way we might setup a series of interactions for an experiment. Then, we are going to calculate and compare the distribution of royalty payout resulting from these varied scenarios under both the total pool and per user payout methods.

What we are going to see is that diversity in listener hours, mixed with that in demand, become the driving forces for any differences in payout outcomes only when those hours and demand are systematically linked with artist popularity. In other words, we will see differences in payouts when comparing the total pool and per user methods only when users who listen to greater amounts of music also listen to a different population of artists than that population listened to by those users who listen to lesser amounts of music.

Therein lays the rub, as far as expectations for the fair and the unfair from any switch in royalty payout method:

If the fans who listen to the greater amounts of music each month also listen to the most popular artists, then it is the less popular—in fact the least popular artists—who benefit most from a switch towards a model that pays royalties on a per-user as opposed to a total pool basis.

If, however, the users who listen to lesser amounts of music each month happen to listen to the most popular artists—a usage pattern various sources of listening data suggest—then, perhaps non-intuitively, it is these popular artists who benefit most from any switch towards a model that pays royalties on a per-user rather than total pool basis.

Furthermore, the so-called “middle class” in any scenario seems to gain the least from any shift between payout methods. Instead, the extremes in the distribution—the most and least popular or demanded artists—see the most notable gains or losses.

Beyond the plusses and the minuses, I hope we are about to see how either of these approaches to royalty payouts can lead to seemingly fair or unfair outcomes. And so, while much of the debate over the distribution of streaming royalties has focused upon the apparent winners and losers, it may well be time to begin to speak more openly about our objectives for these services, and just how we might best play the payout game given these objectives.

The Setup

Before we go any further, lets clarify these two methods for royalty payouts as well as define, at least for the purposes of this work, what we mean by popularity, demand, and listening hours.

The “total pool” approach to royalty payouts aggregates all of the dollars paid by subscribers into one big pool of money, and then pays royalties to each artist/label based upon that artist’s/label’s share of all tracks played across all users. In contrast, the “per user” approach to payouts pays royalties from each, individual user’s subscription dollars to any artist/label based upon that artist’s/label’s share of each, individual user’s listened tracks.

Importantly, both payout approaches involve a “pro rata” (i.e., “in proportion”) consideration. The total pool approach simply considers that proportion of the pool of dollars and plays from the entire population of users, while the per user approach considers that proportion of the dollars and plays from each, individual user account.

As far as the three terms that matter most to this analysis:

Popularity is a measure of the proportion of user playlists in which any artist’s music might be found. Highly popular artists would be found in a greater proportion of user playlists, while lowly popular artists would be found in lesser proportions of user playlists.

Demand is a measure of the proportion of the total tracks played, whether on a per-user or across-all-users basis. Music from high demand artists occupies a greater proportion of total tracks played, while music from low demand artists occupies a lesser proportion of the total tracks played.

Note that the definitions above mean that we can have very popular artists who are in low demand as well lowly popular artists with quite high demand, especially on a per user basis.

Listening Hours is a measure of the time any user spends listening to music. It should come as no surprise, however, that the greater the amount of time people spend listening to music the greater number of tracks they are likely to play.

Each of the following scenarios will be setup as follows:

A set of Subscribers (labeled Q through Z) listens to music from a set of Artists (labeled A through K). You might think of these Subscribers or these Artists as individuals or, ideally, as representative groups—such as deciles from some larger distribution. On average, Subscribers will spend about 111 minutes listening to music each day. The total pool of royalties available will be $7 per Subscriber. As a result, the music will play and the chips will fall where they may.

The Scenarios

Each scenario on the following pages will introduce a different interaction of Popularity, Demand, and Listening Hours, by creating variation both within and among these variables across a set of Subscribers who listen to music created by a set of Artists. Each scenario will also result two distributions of royalty payouts to these Artists, one distribution resulting from a Total Pool approach and the other resulting from a Per User approach to the calculation of these payouts.

The purpose of the effort here is not to advocate for any particular approach—Total Pool or Per User—but to tease out the underlying factors that drive any difference in the distribution of payouts between the Total Pool and Per User methods.

After that, the choice is yours.

Scenario One: You have to start somewhere

Any scenario-based approach to understanding has to start somewhere. Here, we start with the simplest combination of the variables in question: No variation at all. Each Subscriber listens to the same amount of music (900 plays), and each Artist is not only equally popular (all Subscribers listen to each Artist), but also equally in demand (i.e., each Artist sees the same 90 plays from each of the Subscribers, for 900 plays in total).

Untitled

Perhaps as no surpise, therefore, the difference in royalty payout outcomes for the artists in involved does not vary whether we use the Total Pool or Per User approach. Under the Total Pool method, each Artist has 10% of the total pool of plays, and is rewarded with 10% of the royalty pool. Under the Per User Approach, each Artist has 10% of the plays of each Subscriber, leading to a net 10% of the total payouts.

Untitled

As we said, you have to start somewhere. Sort of like a mic check, now we know this thing is on.

Scenario Two: Equal Popularity, Varied Demand, Equal Listening Hours

In this second scenario, we will simply introduce some variation in Demand—the proportion of plays any Artist earns from Each Subscriber, as well as across all plays. Beyond that, each Artist will be equally as popular, and each Subscriber will spend an equal amount of time listening to music.

Untitled

Here we find similarly reassuring results. Artist K, the most in-demand Artists in the crew receives not only the largest payout regardless of payout calculation method, but also there is not difference between these two payouts—whether in raw dollars, or general proportions of the total royalties paid. Artist A, while equally popular is not so in demand, and, therefore, receives the smallest share of the royalty dollars, regardless of method.

Untitled

And so, simply enjoying a greater number of track plays as compared the plays of others is not enough to trigger a difference in payout between the Total Pool and Per User methods.

Scenario Three: Equal Popularity, Varied Demand, Varied Listening Hours

In this scenario, we are going to mix variation in demand with variation in listening hours, while keeping all Artists equally popular. Essentially, while all Subscribers will prefer some Artists over others, some Subscribers will also spend more time listening to music each week, leading to a greater number of tracks played across all Artists.

Untitled

What we see when we turn to the payouts may be unexpected: even when we introduce an extreme variation in listening hours (e.g., Subscriber Z spends almost 100 times the, um, time that Subscribe Q spends listening to music), the royalty payouts across the two methods do not differ—whether in dollar amounts or relative percentages.

Untitled

Apparently, the combination of demand and listening hours is not enough to shift apart the payouts from these approaches. We are going to have to alter popularity.

Scenario Four: Slightly Varied Popularity, Varied Demand, Varied Listening Hours

In our fourth scenario we are going to begin to mix variation in popularity with that in listening hours and demand. Two of the Artists, played by only nine of the Subscribers, will be only slightly less popular than the remaining eight Artists. However, zeros matter.

Untitled

Turning to the payouts, we begin to see some deviation between the Total Pool and Per User methods. While slight in dollar terms, these differences are measurable in percentage terms. Artist K (the most in demand among the less popular artists) earns 13.6% less under the Per User approach, Artist A (the least in the demand among the less popular artists) earns nearly 50% more, while the Artists in “the middle” each see a small bonus of 1.76%.

Untitled

Interesting.

Scenario Five: Some Variation in Popularity and Distinct Populations, Varied Demand, Varied Listening Hours

In Scenario Five, we further twist the variation introduced in the prior scenario. We have variation in demand as well as listening hours. Not to mention, we have some variation in Popularity, with some Artists having a five, six, and seven Subscribers and on with eight Subscribers. Importantly, however, the Subscriber populations who select these Artists are different: those Subscribers who spend the least amount of time listening to music each month select different artists from those Subscribers who listen to the greatest amounts of music.

Untitled

When we turn to the payouts we see more clearly how listening hours, demand, and popularity interact in order to drive an difference in outcomes between the Total Pool and Per User payout methods. Artist K, the most demanded artists from the voracious listeners, sees a 54% decline in payout from the Per User approach, while Artist A, the least demanded artist by the least voracious listeners sees a 1000% increase in payout.

Untitled

Whoa, that’s a redistribution of dollars.

However, all Artists aren’t equally popular in the music world in which we actually live. Not to mention, we still have to pursue additional scenarios in order to dig deeper into what might seem fair or unfair.

In the next post, we are going to introduce variation in Popularity across the Artists so that we can see how the interaction of all of these variable can lead to outcomes quite different from what might be expected.

Redistributing the streaming wealth: Why listening hours not artist popularity may drive the divvying up of the royalty pot

The question of just how to distribute the royalty dollars from music services has moved from a back-office thought exercise to an out-in-the-open debate.

Moving forward in this debate is the argument that it would be fairer—particularly to emerging, independent, or less “popular” artists—to distribute royalties on an each pair of ears basis (i.e., per subscriber) rather than on an all pairs of ears basis (i.s., all subscribers). See Laguana’s “How to make streaing royalties fairer,” and Gottfried’s, “Throw out the big pot: A re-examination of streaming pay” (largely citing Jeremy Silver) as examples of this debate.

Unfortunately, yet perhaps non-intuitively, it may only be the case that these artists, or any artist for that matter, would benefit from a per pair of ears approach to royalty payouts if it is also true that people who listen to these artists happen to spend less time listening to music. People may be confusing what happens when an artist is popular or not (i.e., a large or small number of people listen to that artist) with what happens when listeners spend a greater or a lesser amount of time listening to music.

More specifically, I wonder if folks may be overlooking the fact that what drives the variation in royalty payouts from music services that have adopted a “percentage of listens” (or, pro rata) approach to these payouts is the diversity in listening hours across listeners, leading to differences in the number of tracks streamed, rather than some diversity in artist selections across listeners—a situation that may or may not impact the time spent listening to any number of tracks.

Simply stated: how people are listening matters regardless of to whom they are listening, when the question on the table is the implied price or value of royalty payouts from music services.

My ultimate concern about this debate is that the each pair of ears perspective may be far to focused upon the trees—as in, the per stream payouts—at the expense of the forest—as in, how listening behavior across listeners drives the total value of payouts. That may be a mixed metaphor…

In this post, we will use some simple scenarios to highlight what I reckon to be the different impacts upon royalty payouts that result from diversity in listening hours versus those from diversity in artist selection. As a result, I hope readers will better understand the drivers of royalty payouts in a pro-rata world.

My apologies for the length of this post, but it would seem that digging into the details of this payout dilemma takes a greater number of words than can fit into a 140 character tweet, or a blurb in the blogosphere.

The ears have it

The great thing about how streaming services pay music stakeholders is that actual listening behavior drives compensation. That said, there is a difference of opinion over how the ears should matter.

Summarizing this difference of opinion regarding the each-versus-all pair(s) of ears perspectives:

Per pair of ear basis
Split the $X royalty pool resulting from each subscriber’s payment according to that subscriber’s listening behavior.

For example, if Jack listened to only a single artist last month, while Jill listened to ten artists, pay all of the royalty dollars from Jack’s subscription to his artist and all of the royalties dollars from Jill’s subscription to her artists.

All pairs of ears basis
Aggregate the $X from all subscribers into a single, royalty pool, and split these dollars according to the aggregated listening patterns of all subscribers.

For example, if Jack listened to a single artist and Jill listened to ten artists, pay these eleven artists out of the total royalty pool resulting from the combination of the dollars from both Jack’s and Jill’s subscriptions, according to that percentage of all listening attributed to each of these artist’s work.

As I claimed earlier, diversity in listening hours leading to variations in the number of tracks streamed—or, how people listen to music—drives the royalty payouts regardless of diversity in artist selection—or, to whom people are listening. Let’s use Jack and Jill to explain this seemingly erroroneous claim.

Meet Jack and Jill

Jack and Jill are subscribers to a music service we will call Noisebox. What can I say, I still like that name.

Jack and Jill each pay $1 per month for access to this Noisebox music service—a bargain, but this example is supposed to be simple.

Jack is a music Monovore: He listens to only one artist. Ever.
Jill is a music Polyvore: She listens to many artists. Always.
Jack and Jill both like Artist A.

The entirety of subscription dollars collected by Noisebox are paid through as royalties to artists—a financial implausibility, but again, this is supposed to be a simple example.

In our simple world, the artists behind the music licensed by Noisebox are not just the featured artists, but also happen to be the writers, publishers, labels, producers, performing musicians, and whoever/whatever else might have a stake in the recordings. And, so we only need to pay royalties to these artists. Dealing with payments to all parties in a multi-stakeholder world is a topic for a book not a blog post.

Noisebox uses a “percentage of listens” (or pro rata) approach to distributing the royalties that result from Jack’s and Jill’s payment for and use of the music service. Which means, we simply take the dollar on the table, divide that dollar by the number of listens, and distribute these smaller sums according to the artists behind the songs.

For example, if there were ten listens in the month, we simply divide $1 by ten, the result being $0.10 (ten cents) to be paid for each listen. Should all listening have been to the music of one artist, that artist would receive ten payments of $0.10. Should a different artist be associated with each of these ten listens, then each of these artists would receive a payment of $0.10.

Now let’s let Jack and Jill listen to some music, so Noisebox can start paying some royalties. The only challenge being, Jack and Jill are not like each other, the result being that how and to whom they listen is quite different.

Let the music play:
Same time spent, different artists

In this scenario, both Jack and Jill spend enough time listening to music this month that they each stream ten songs. However, Jack streams ten songs from the same artist (artist A), while Jill streams ten songs from ten different artists (artists A, B, C, D, E, F, G, H, I, and J — but not JayZ, he has his own music service). And, we have two dollars on the table for royalties, as both Jack and Jill paid Noisebox $1 this month for the music.

Let’s looks at the royalty payout outcomes for this scenario under an all pairs of ears as well as an each pair of ears approach.

All pairs of ears
In this approach, we aggregate all dollars and then pay royalties based upon each and any artist’s percentage of all aggregated listens.

20 Total Plays = ten plays from Jack plus ten plays from Jill
$0.10 per play, royalty = A $2 pool of dollars, split twenty ways

Artist payouts:

$1.10 => A (11 out of 20 plays)
$0.10 => B (1 out of 20 plays)
$0.10 => C
$0.10 => D
$0.10 => E
$0.10 => F
$0.10 => G
$0.10 => H
$0.10 => I
$0.10 => J

Each pair of ears
In this approach, we take the dollars from Jack and pay them as royalties according the each artist’s percentage of Jack’s listening. We apply the same method for Jill and paying her artists.

Payouts from Jack:
$0.10 per play, royalty = $1 divided by ten plays

Payouts from Jill:
$0.10 per play, royalty = $1 divided by ten plays

Artist payouts:
$1.10 => A (ten of Jack’s ten plays, plus one of Jill’s ten plays)
$0.10 => B
$0.10 => C
$0.10 => D
$0.10 => E
$0.10 => F
$0.10 => G
$0.10 => H
$0.10 => I
$0.10 => J

NOTE: Even though the selection of artists varied greatly between our two subscribers, Jack and Jill, the royalty payouts did not differ for the artists involved. In particular, Artist A—as popular as before—saw the same payout across both approaches, even though one fan is exclusive to A, while the other is non-exclusive. Importantly, the example above scales even if we add subscribers who never listen to Artist A. Add 50 more subscribers, none of whom listen to Artist A, but all of whom listen to ten tracks, and you will find that Artist A receives the same payout: $1.10.

Let the music play:
Different time spent, different artists

Now let’s look at a couple scenarios in which Jack and Jill differ in terms of the amount of time they spend listening to music, therefore streaming a different number of tracks.

In the first scenario, Jack will only listen to Artist A once, having spent only a few minutes using Noisebox, while Jill’s listening stays the same as before. This is the scenario many people are actually describing when they imagine the impacts of the each pair of ears approach—Jack not only is an Omnivore, but hardly even listens to music.

In the second scenario, Jill will listen to each of her ten artists twice, the result of spending twice as much time using Noisebox, while Jack’s listening will remain the same as above.

Scenario One

All pairs of ears, Jack listens to Artist A, only once.

11 Total Plays = one play from Jack plus ten plays from Jill
$0.1818 per play, royalty = A $2 pool of dollars, split eleven ways

Artist payouts:

$0.3636 => A (2 out of 11 plays)
$0.1818 => B (1 out of 11 plays)
$0.1818 => C
$0.1818 => D
$0.1818 => E
$0.1818 => F
$0.1818 => G
$0.1818 => H
$0.1818 => I
$0.1818 => J

Each pair of ears, Jack listens to Artist A, only once.

$1 per play, royalty from Jack = A $1 pool of dollars, split one way
$0.10 per play, royalty from Jill = A $1 pool of dollars, split ten ways

Artist payouts:

$1.10 => A (All of Jack’s payment, plus one of )
$0.10 => B (1 out of 20 plays)
$0.10 => C
$0.10 => D
$0.10 => E
$0.10 => F
$0.10 => G
$0.10 => H
$0.10 => I
$0.10 => J

NOTE: The diversity of artists to which Jack and Jill listened stayed the same in this example. Furthermore, Artist A was a popular as before. What drove the difference in payouts to Artist A between the two methods was simply the fact that Jack spent less time listening to music. In this case, if we add 50 more subscribers who never listen to Artist A, but all of whom listen to ten tracks, you will find that Artist A receives the $1.10 payout only in the each pair of ears approach. What is driving this difference in payouts — the time Jack spent listening to music, not the diversity of artists.

Scenario Two

All pairs of ears, Jill listens to twice as much music

30 Total Plays = ten plays from Jack plus twenty plays from Jill
$0.0667 per play, royalty = A $2 pool of dollars, split 30 ways

Artist payouts:

$0.1333 => A (2 out of 30 plays)
$0.0667 => B (1 out of 30 plays)
$0.0667 => C
$0.0667 => D
$0.0667 => E
$0.0667 => F
$0.0667 => G
$0.0667 => H
$0.0667 => I
$0.0667 => J

Each pair of ears, Jill listens to twice as much music

$0.10 per play, royalty from Jack = A $1 pool of dollars, split ten way
$0.05 per play, royalty from Jill = A $1 pool of dollars, split twenty ways

Artist payouts:

$1.05 => A (All of Jack’s payment, plus one of Jill’s payouts)
$0.10 => B (2 out of 30 plays)
$0.10 => C
$0.10 => D
$0.10 => E
$0.10 => F
$0.10 => G
$0.10 => H
$0.10 => I
$0.10 => J

NOTE: The diversity of artists to which Jack and Jill listened stayed the same in second scenario. And again, Artist A was just as popular as before (being in both Jack’s and Jill’s streams). What drove the difference in payouts to Artist A between the two methods was simply the fact that Jill spent more time listening to music. The role of listening time may now be more clear. What drove the difference in payout for Artist A under the each pair of ears approach across the two scenarios was Jill’s listening behavior, not Jack’s.

The end

That was one heck of a post to support a very simple claim:

How people are listening matters regardless of to whom they are listening, when the question on the table is the implied price or value of royalty payouts from pro rata music services. An artist being popular or not (i.e., a large or small number of people listen to that artist) may have a very different effect as compared to that of people spending a greater or a lesser amount of time listening to music.

0.02%: the possible Sales-to-Spins conversion rate of US Radio

It has often been said, I don’t know by whom, that the power of Radio is it’s ability to sell music. The purpose of this post is to surface any possible evidence of, if not simply a measure of, this claim.

All we are going to do in this post is measure the Sales-to-Spins conversion rate of US Radio performances in the same way that online advertising is measured — the number of music sales transactions that occurred in 2014 as compared to the number of music performances, or  “impressions,” across all listeners of Radio in the US of A.

What we will find:

The maximum Sales-to-Spins conversion rate of US Radio may be no greater than: 0.1%.

(Remember, this estimate makes the horrifically implausible assumption that the only trigger for a music sale is a Radio play.)

And, if we believe the Radio industry’s own study, which claims that 14-23% of sales are triggered by Radio play, then:

The minimum Sales-to-Spins conversion rate of US Radio may be as low as: 0.02%

Alternatively stated: No greater than 1 in 1000 performances to some listener convert to a music sales transaction of some kind (i.e., a downloaded single, an album sale, etc.). Or, as few as 2 in 10,000 Radio performances to some listener may convert to a music sales transaction.

In estimating this ratio, all we did was take the total number of music sales transactions in 2014, as reported by the RIAA:

1,516,650,000 music sales transactions

And divide that number by some estimate of the total number of music performances on US Radio in 2014, across all listeners—the closest equivalent to “impressions” (which are simply the number of views of an ad by all viewers):

1,485,098,654,916 performances to listeners on US Radio

1,516,650,000 / 1,485,098,654,916 is simply: 0.10212453%

The above is the maximum estimate, the unstated assumption within which being that all possible music sales transaction might be attribute to performances on the Radio.

Given the above assumption may be a bit, aggressive, we can adjust the number of music sales transaction by the percentage of those music sales transactions attributed Radio, by the National Association of Broadcasters (a 2008 study, released by the NAB itself):

280,580,250 music sales transactions

And divide that number by the estimate of music performances on US Radio in 2014:

280,580,250 / 1,485,098,654,916 is simply: 0.018893%.

Or, 2 in 10,000 performances.

If anyone would like to check the maths, all you have to do is follow this link.

 

Earth to the US Music Industry: 99.9% of your revenues now come from Digital sources

Let’s just be honest. If there were any industry that prefers to deny reality it is the music industry.

There may be no greater evidence of the claim above than the following statement from Cary Sherman, Chairman and CEO of the RIAA:

“The music business continues to undergo a staggering transformation, one embraced by the music labels we represent. Record companies are now digital music firms, earning more than 2/3rds of their revenues from a variety of digital formats.” (the link in the press release heads to Why Music Matters, for no apparent reason).

Earth the RIAA: The CD is a digital format!

If you disagree, please place your favorite CD on an analog turntable, place the needle on the CD, and let us know what you hear.

Of the revenue sources listed in the RIAA’s Year-End Shipment and Revenue Statistics, only one source of revenue stands out as being truly Analog: the Vinyl Single, worth $5.9 million of revenue in 2014, or 0.084621783655% of revenue.

100% – 0.084621783655% = 99.915378216345%

 

If we grant some analog reprieve to the LP/EP (worth $314.9 million in revenue), then it would seem that at best, $320.8 million of the $6.9722 billion in 2014 revenue would come from digital sources—or, 4.601130202805%

100% – 4.601130202805% = 95.398869797195%, or 95.4%

But, let’s face it, even the majority of the LP/EP category is now a digital file—music packaged as 1’s and 0’s, on a shiny disc.

And so, 99.9% of US Music Industry revenues now come from Digital sources.

Feel free to disagree.

Math-checking the testimony on Capital Hill in “How Much For A Song?: The Antitrust Decrees That Govern the Market for Music.”

According to Billboard, SONGS Publishing CEO Matt Pincus provided some mathematical evidence before members of Congress today in support of his claim that “This rate of money [from performances on Pandora] is not fair for my songwriters.”

Now, I could and would never take a stand on what royalty rates might be “fair” when it comes to the fates of songwriters, publishers, labels, and performing artists. What I can do, however, is use data to check the maths that emerge from these arguments—particularly arguments made in testimony on Capital Hill—both for and against whatever might be fair. We can as least try to debate the fairness of the same numbers!

In this testimony, Mr. Pincus referred to three songwriters affiliated with SONGS Publishing receiving only $3,158 for 124,000,000 streams on Pandora. Unfortunately, this amount may not tell the entire songwriter story.

In this post, we are going to compare estimates based upon claims of the Payee to those based on the claims of the Payor, and see where the numbers fall.

To do this comparison, we are going to adjust Mr. Pincus’ figure for the fact that there may actually be six writers, not only three, for the song in question—Kama Sutra, performed by Jason Derulo (perhaps also Kid Ink)—suggesting the total amount paid to songwriters might be twice the amount referenced in testimony.

Then we will add to this revised songwriter payout that amount paid to publishers, as well as the administrative fees charged by related PRO(s).

After which, we are going to check those maths by estimating an effective per stream payout to songwriters and publishers and PROs, per stream, from Pandora; using Pandora’s own, publicly available financials.

Payee: What we are going to find is that if we adjust Mr. Pincus’ numbers for what wasn’t said, we will estimate a rate of $0.00012 per stream paid to songwriters/publishers/PROs on a per stream basis via Pandora.

Payor: And, if we use Pandora’s own financial statements, we will estimate a rate of $0.00012 per stream paid to songwriters/publishers/PROs on a per stream basis.

After which, the planets align, the oceans return to safer levels, and the debate over fairness can continue using, at the very least, the same numbers.

The Deep Dive

Now, in the following exercise I may have done my math wrong. But, at least I will show you my math.

While the exact song to which Mr. Pincus referred was unclear, a quick search surfaces the result that SONGS publishing has a claim in the song Kama Sutra, performed by Jason Derulo (and also Kid Ink). Feel free to investigate AllMusic or other sources to confirm.

Six writers are associated with Kama Sutra: Brian Collins, Christian Ward, Jason Desrouleaux, Dijon McFarlane, Mikely Adam, and Breyan Isaac.

In testimony, Mr. Pincus refers to $3,158.05 as the payout made to three writers. If there are in fact six writers, a simple back-of-napkin adjustment would be to double the figure, leading to a total of $6,316.10 paid to the six writers of Kama Sutra for these millions of performances on Pandora.

Now, Sony/ATV, Warner/Chappell, Universal Music, and SONGS publishing are listed as the publishers associated with the song. Let’s just assume a 50/50 split between songwriters and publishers in the royalties received for the performance of the song. Not a crazy assumption given these are “blue chip” publishers, and we are talking about payments through the PRO’s, and Mr. Pincus may have labeled these amounts as the writers’ 50% share.

Such a move would lead us to $12,632.20 being the total payout for these 124,000,000 performances on Pandora: $6316.10 to songwriters and $6316.10 to the publishers.

Since these royalties would have been paid through a US Performance Rights Organization, let’s say ASCAP (since Pincus is on an ASCAP Board), some fees would have been taken by the PRO—approximately 12% of total song royalties using ASCAP’s latest figures. As a result, the total paid for the performance of the song would be approximately $14,354.77.

Now $14,354.77 divided by 124,000,000 streams on Pandora would result in an effective payout of approximately $0.00012 per stream. Or, if you would rather be extremely precise, $0.00011576429619 per stream.

So let’s compare that figure to what we might estimate using Pandora’s own numbers.

Pandora’s end-of-2014 financials we learn the company earned roughly $920,800,000 in revenue in 2014. And, there is reason to believe, at least according to sources like the LA Times referencing BMI’s own testimony in a rate-setting debate, that 4% of Pandora’s revenue is paid to US PROs for the performance of songs (not recordings).

Four percent of $920.8 is roughly $36,832,000 paid in royalties to the PROs, roughly 88% of which passes through to songwriters and publishers.

We also learn that the company streamed 20,030,000,000 (that’s 20.03 billion) listener hours in 2014. And, using the very back-of-napkin assumption of songs with an average length of just under four minutes, we guesstimate about 15 songs played per hour with some time left over for ads.

20.03 billion hours of listening at 15 songs per hour leads to 300,450,000,000 unique song streams on Pandora in 2014.

$36,832,000 in total payouts divided by 300,450,000,000 streams equals an effective payout of approximately $0.00012 per stream.  Or, if you would rather be very precise with guesstimates, $0.00012258944916 per stream — an amount that differs from the prior estimate by 5%.

Given we are working with numbers that extend from the millions of dollars and billions of uses to the fractions of pennies, I am inclined to find a 5% difference in two estimates to be within a reasonable margin of error.

And so, we find that if we adjust Mr. Pincus’ numbers for what may not have been said, we will estimate a rate of $0.00012 per stream paid to songwriters/publishers/PROs on a per stream basis via Pandora.

And, if we use Pandora’s own financial statements, we will estimate a rate of $0.00012 per stream paid to songwriters/publishers/PROs on a per stream basis.

After which, the planets align, the oceans return to safer levels, and the debate over fairness can continue using, at the very least, the same numbers.