News

Just how big is the music industry’s ‘fake streams’ problem?


Tags:

The music industry has been fretting about fraudulent streams for several years now. Our first mention of ‘fake streams’ came in July 2017, and a couple of years later we reported on the launch of an ‘Anti-Stream Manipulation Code of Conduct’ signed by labels, publishers and DSPs.

Since then, the grumbling about how bad actors are trying (and more importantly, succeeding) in gaming the streaming system and siphoning royalties away from artists and rightsholders has slowly but surely grown in volume.

“Some bad actors who do not share our commitment to artists and artistry have been swooping into the reinvigorated industry,” is how UMG boss Sir Lucian Grainge put it this month. “The current environment has attracted players who see an economic opportunity in flooding platforms with all sorts of irrelevant content that deprives both artists and labels from the compensation they deserve.”

With that in mind, a new study published by the Centre Nationale de la Musique (CNM) in France is already making waves within the industry. Based on data supplied by Spotify, Deezer and Qobuz as well as labels and distributors (all three majors, Believe and Wagram) the study estimates that between 1% and 3% of total streams in France in 2021 were ‘false’. That’s 1-3bn streams.

One important note: these are false streams detected by the streaming services and “éliminés du partage des droits” (eliminated from their sharing of rights), so the CNM is sure that the real number of fake streams is higher.

Recent research published by Luminate allows us to scale this up globally. It said that there were 3.4 trillion on-demand audio streams worldwide in 2022. If we even just apply the conservative 1-3% estimate to that, it could mean between 34bn and 102bn fraudulent streams annually.

(MBW, which reported on the study yesterday, applied the figures to global revenues, using the IFPI’s stat of $16.9bn of trade revenues from streaming services in 2021. 1% of that would be $169m while 3% would be $507m.)

The first question: how big a problem is this, if upwards of 97% of streams are non-fraudulent? In thinking about this question, we found our way to a European Central Bank study on [bank] card fraud. In 2019, fraud was just 0.036% of overall card transactions worldwide.

That may suggest a target for DSPs in our world, although let’s keep in mind that the modern streaming business is barely 15 years old – a baby in terms of both experience and resources of tackling fraud compared to the financial industry.

A second question: are streaming services taking this challenge seriously? The CNM has pointed the finger in this regard, noting in its announcement that Amazon Music, Apple Music and YouTube “n’aient pu ou souhaité partager leurs données suivant le périmètre d’observation défini, malgré toutes les garanties de confidentialité que le CNM leur apportait” (were unable or unwilling to share their data according to the parameters of its study, despite its guarantee of confidentiality).

That seems disappointing, although those three services may have their reasons and perspectives on the CNM’s methodology and guarantee. You can imagine rightsholders will be asking about that behind closed doors, if not in public.

What happens next? The CNM noted that while there are current legal frameworks to tackle streaming fraud, it can be difficult to identify the criminals. Instead, it called for more action from the music industry, including a charter defining streaming manipulation, and formalising and harmonising “des processus d’alerte, ainsi que des sanctions graduées” (alert processes, and also graduated sanctions).

A code of conduct, y’say? Like the one signed by many of these industry players (but not Apple or YouTube) back in 2019? If we remember this bit of industry history, at least we won’t be doomed to merely repeat it.

The key this time round will be to move beyond saying the right words – that streaming manipulation is bad – and to detailed, constructive discussions not just about how current methods can be tackled, but about what kind of technologies, expertise and investment will be required for streaming services to be prepared for future fraudulent schemes too.


Written by: Stuart Dredge