Tears in Rain

Do Androids Stream Electric Sleep?

There were two bits of music industry news last week that I’d like to connect, because I believe they point to widespread fraud.

The first is an amusing, hustler-games-the-system story from Rolling Stone. Stef Van Vugt, aka Steve Void, is a Dutch DJ and producer who runs a label called Strange Fruits – which is busting Spotify charts with 30-second snippets of rain. The idea of sleep-enhancing playlists isn’t his invention, nor is the idea of breaking up recordings into 30-second tracks in order to increase click counts (30 seconds is the minimum to register a paid stream). But Van Vugt has reached global top ten numbers on Spotify by spending, he says, “over eight figures” on marketing. This has earned his sounds of rain a Lady Gaga-scale streaming income, as Rolling Stone documents.

I wouldn’t take Van Vugt’s boast about his marketing budget at face value – there’s no way to know if he has literally paid tens of millions in his quest for Spotify riches. But his streaming numbers are verifiable. Van Vugt is proving that throwing money at streaming can earn back even more.

How, exactly? You can’t book rain on late night TV, or get it headline slots at festivals.

A second news item points toward an answer. This one is a hopeful, we-can-fix-the-system story from Pitchfork: over on SoundCloud, switching their accounting for streams from the prevailing “pro-rata” system to a “user-centric” method resulted in a 500% increase of income for a track by Portishead. On the same number of streams.

Now from personal experience, I know that “pro-rata” and “user-centric” are the kinds of terms that cause people to stop listening to whatever you are saying. In my advocacy work with UMAW, my comrades and I on the Streaming Committee have learned this is the sort of insider talk that cannot survive in the more robust air outside industry circles. What difference could it possibly make, which way you count up streams?

I’m going to take advantage of this subscription service to write about the topic anyway. Cause I swear, it’s important for music.

User-centric accounting is what you probably assume is happening when you stream a song; that is, the song earns that particular artist a portion of the money you put into the system (whether by subscription or by listening to ads). But that’s not how it’s set up on Spotify, Apple Music, or anywhere else at the moment other than a part of SoundCloud that includes the Portishead song. What happens instead is referred to as pro-rata accounting: your stream registers as a click toward that artist’s share of all streaming income, which is then divided up according to relative total usage.

Those two ways of accounting sound like they might be the same: whatever you listen to gets some money, and Drake gets the rest. Six of one, half dozen of the other.

But if that were the case, how did Portishead earn six times the amount from the same streams accounted for one way (user-centric) rather than the other (pro-rata)? There’s clearly something missing in the equation. And what’s missing may well be Stef Van Vugt’s sounds of rain.

Because where did all that money in Van Vugt’s marketing budget go? Spotify do offer versions of what is essentially payola, taking money in exchange for algorithmic favoritism – but that can’t explain this extreme example. What would explain it, though, is if you could buy clicks that simply pay back more than they cost. Then the more you spend, the more you earn. Since a subscription buys unlimited plays, a bot might use a single account to register clicks for an inhuman amount of listening – skewing the percentage pulled out of the collective pot created by pro-rata accounting. Real listeners are lost in the flood.

Indeed, if you google “Buy Spotify plays,” you’ll find any number of offers that are variations of this one:

“Starting at 1,000 plays for $4 and going up to 500,000 plays for $745, you can take a pick according to your budget and needs.”

I have no idea which if any of these offers are real and which are scams, but the going rate does seem logical, at least as a come-on. At the lower end, you lose money ($4 for 1,000 streams on Spotify will earn back about $3 in royalties). And at the higher end, you double your money ($745 for 500,000 streams will earn back about $1500). Does it really work? I’m not going to spend any cash to find out.

But let’s say it does – and some variation of click farms is what Van Vugt successfully uses to boost his playlists of rain to rival even Spotify’s most popular ones like Rap Caviar. Wouldn’t that explain the difference between what Portishead’s song earned from user-centric accounting, which registered actual listeners to the track, vs pro-rata, which credited the band with a proportional share of all streams, including any fraudulent ones?

On to the bigger question: Why wouldn’t Spotify leap to eliminate fraud like this, which sends their money to hustlers like Stef Van Vugt? User-centric accounting neutralizes bots, because regardless of the number of plays, each listener can direct only the full value of one subscription toward any given artist - Van Vugt would have to spend $9.99 to get his rain credited with $9.99. Yet here’s how Spotify addresses the question on its “Loud and Clear” website, intended to counteract critiques of their system:

“A shift to user-centric payments would not benefit artists as much as many may have originally hoped… the change would result in ‘at most a few euros per year on average’ for artists outside the top 10,000.” [Italics mine]

Slipping in the qualifier “outside the top 10,000” is key to that statement, which sidesteps the question of what the switch would do to the top earners from streaming, and to the platform itself. If you understand that user-centric accounting might significantly reduce earnings for the winners of the current system, Spotify’s next remark may read differently too:

“We are willing to make the switch to a user-centric model… However, Spotify cannot make this decision on its own – it requires broad industry alignment to implement this change.”

Translation: we’re never going to switch to user-centric unless we’re absolutely forced to. It seems that Spotify - and their major label partners, who helped create this system and eagerly support it - are just as invested in pro-rata accounting as Stef Van Vugt.

In other words, they benefit from the same type of fraud.

The mechanism by which they benefit is easy to understand: for Spotify, inflated streaming numbers means inflated market share, which ultimately means real cash. Each quarter, Spotify reports a loss. Meanwhile, their increasing market dominance of streaming (32%, double their nearest competitor Apple at 16%), and streaming’s increasing dominance of the industry as a whole (84% of all recorded music revenue), has earned them a stock market capitalization currently valued at $47 billion. Fortunes have been made, and are being made, from this business run in the red.

What if user-centric accounting resulted in a huge reduction of reported streaming numbers by limiting them to actual streams listened to by actual people? What if Spotify’s total numbers were off by as much as 500%, as Portishead’s experiment on SoundCloud suggests is possible, and the top earners’ streaming numbers were as well? Stef Van Vugt would be out a lot of money, but so would Spotify, and so would the three major labels – who control a disproportionate percentage of music on the streaming charts.

About those charts… The last time they were shaken by a profound data correction was May 1991, when Billboard started using SoundScan to count actual record sales to actual customers, rather than relying on inflated estimates filed by store managers and distribution execs (whose entire business depended on favorable transactions with the major labels). Turned out people were not buying records in precise proportion to the size of billboards on Sunset Blvd, but had all kinds of undercounted enthusiasm for country, hip-hop, speedmetal, and what became known as “alternative.” Over the next six months, Garth Brooks, NWA, Metallica and Nirvana unexpectedly hit number one in place of any number of pop acts that the industry had planned for instead.

Those same charts were altered again in December 2014, this time to add a formula for estimating streaming numbers as “album equivalents.” “The most substantial methodology update since May 1991,” said Billboard in their announcement.

If widespread fraud is manipulating streaming counts – as I believe it is – then we are back in a pre-SoundScan era of inflated numbers boosting false demand for a limited range of music which is profiting wildly, while the rest of us struggle to attach any monetary value to our recordings. All our music is just tears in rain.

Listening to: Hassan Wargui, Tiddukla

Cooking: Tarhana (dried wheat and yogurt) simmered with green tomatoes and sorrel