Enter the World of Streaming Manipulation, Part 1 - Fake Playlists

Fake streams! Playlist manipulation! Fake artists! There’s a lot of buzz about it, but what does this look like in the data? ......chartmetric.com @chartmetric #chartmetric #musicanalysis #musicbusiness #fakestreams #fakeartists #fakeplaylists #musicbusinessworldwide


  • Highlights  
    • Fake streams! Playlist manipulation! Fake artists! There’s a lot of buzz about it, but what does this look like in the data?
  • Mission   
    • Good morning, it’s Jason here at Chartmetric with your 3-minute Data Dump where we upload charts, artists and playlists into your brain so you can stay up on the latest in the music data world.
  • Date
    • This is your Data Dump for Friday, June 28th, 2019.
  • Enter the World of Streaming Manipulation
    • Last week’s streaming code of conduct was signed by more than 20 major companies across the industry to combat streaming fraud, which is good for artist compensation and more forthcoming to the fans.
    • How can we think about this prickly topic from a music data perspective? And when we say “this”, it’s not just fake streams. It’s also playlist manipulation and fake artist accounts.
    • For sure, we are in very murky waters, and there is little actual data on the phenomenon.
      • Recently American indie label Hopeless Records estimated 3-4 percent of global streams could be fraudulent.
      • But a 2015 MBW article mentions how 60% or more Twitter followers on top artist accounts could also be fake.
      • Granted, these are different types of fraudulent behavior, but it’s also a huge delta to try to account for.
      • What we can do though is search for red flags in the music data available to us.
    • For example: if we look at playlist manipulation, here’s one way to look at the data to try to identify potentially iffy behavior:
      • We scanned the playlist charts looking for abnormally high 28-day follower increases, and found a non-editorial hip-hop genre playlist with a 262% increase in followers in the past month.
      • While that could just be great marketing, currently having 110K followers-an impressive number-its max artist monthly listeners, however, is only ~470, which doesn’t seem to match up.
        • This means that the only artist on the playlist that gets a lot of its unique listeners from here is getting less than 1% of its supposed followers actually listening to them.
        • Again, possible, especially since the playlist has about 100 current tracks on it, but it’s ranked in the first third of the playlist, so it’s not likely.
      • That artist, which only has a little over 200 followers, is playlisted among high-profile artists like Eminem, Kanye West and Cardi B, presumably to draw traffic, which would be smart marketing if done legitimately, but if so many followers are not streaming the actual tracks...it smells a little fishy.
      • If that weren’t enough, there’s a three-piece pop band with only 16 followers, and two other rap artists who have 4 and 17 Spotify followers, respectively.
        • All three have their listed label as a series of numbers, then “Records DK” or “DK2”, which is a default label for the distributor DistroKid, if left untouched.
        • DistroKid is one of the most popular digital distributors available to independent artists and an official partner distributor with Spotify.
      • If that still isn’t enough, all the playlist album artwork looks like carbon copies of official Spotify playlist album art. Again, good marketing tactic...or borderline deception?
        • So while it’s admittedly an analytical leap, it is very possible that a playlist curator is buying illegitimate playlist followers to make themselves look good, they dupe unknowing artists into thinking they are getting amazing exposure, and the curator gets paid accordingly and in our opinion, unfairly.
        • We could be completely 100% wrong on this, but the point is, there are certain ways you can look at the music data to try to suss out what’s likely real, and what at least should raise some red flags.
    • We’ll try to unpack some other types of illegitimate activity from a data perspective next week.
Outro
  • That’s it for your Daily Data Dump for Friday, June 28th, 2019. This is Jason from Chartmetric.
  • Do you know how NPR does their ask for donations every so often? That’s what we’re about to do now! But we’re just asking for an Apple Podcasts rating.
  • Rutger and I put at least a few hours a day into each episode, researching, writing, editing, recording, editing again, publishing to multiple platforms, checking analytics...and it’d be really cool for us to get some feedback on how we’re doing: the good/bad/ugly. So it’d only takes a few thumb swipes out of your day, and you’d be sending us so much joy: we’d appreciate it.
  • As always, free accounts are at chartmetric.com
  • And article links and show notes are at: podcast.chartmetric.com
  • Happy Friday, have a great weekend, and see you on Monday!

2019 Chartmetric