Fraudsters have been using bots to artificially boost the popularity of selected music tracks for their own gains.

Every morning, millions of people wake up with a cup of coffee while listening to their favorite music.  There is a very good chance that the track was on a music streaming service.

With the increasing adoption of a digital lifestyle within the region, the Asia Pacific region’s online music streaming market is projected to surpass US$14bn by 2027.

However, as bad actors attempt to prey on streaming services and steal revenue from musicians and artists, the music streaming industry must take definitive steps to win the fight against fraudsters using bots for ‘bulk streams’ and ‘fake listens’.

Attacking the streaming economy

When people subscribe to a streaming service, the accumulated amount is added to a pool. Streaming platforms are given a cut, but the bulk of the earnings goes to labels, distributors, publishers and collecting societies. Musicians are last on the list to get paid based on the terms of their contracts.

Fraudsters can eat into the earnings in two ways:

  1. The first is to play ‘fake’ tracks owned by non-artists on a loop. This allows artists to gain ‘listens’ and to appear more popular with higher rankings and ratings on streaming platforms. Bots allow non-artists to skim millions of dollars away from legitimate artists by creating thousands of fake accounts and playing tracks registered to the fraudster.

    On many streaming services, a track needs to be at least 30 seconds long to be monetizable. With 86,400 seconds in a day, bot outfits can play up to 2,880 tracks a day per device, totaling 2.9 million tracks using 1,000 bots. This comes up to some 86.4 million illegitimate plays (fake listens) in the ballpark of a quarter of a million dollars’ worth of revenue per month.
  2. Unscrupulous bot vendors offer the ability to circumvent the natural order of customers’ behavior of relying on recommended playlists curated through the music service’s algorithms that measure the popularity of tracks for search engine and playlist placement.

    For a simple upfront payment, fraudsters offer the ability to provide a huge bot farm of devices to listen to specific tracks thousands of times (‘bulk streaming’), which translates to higher placement in searches, the possibility to be featured in hot playlists, and most of all—gain listeners’ attention. Vendors can use this ‘bulk stream’ service to game the system by fabricating attention to get attention.

    While this does not have a direct economic hit on artists, it does shift the goalposts. Those who can afford to pay bot vendors get the attention instead of the artists who truly deserve it.

Beating the bots with AI

For both record labels and streaming services, validating the integrity of streams is essential to ensuring equitable allocation of streaming revenues to artists and rights holders.

However, similar to gaming and e-commerce verticals, the threat of bot fraud is ever-present and will continue to be so.  

As protection from fraud becomes more sophisticated, the attackers will also keep up. The key to staying ahead in this game is to stay prepared at an organizational level. 

Businesses should have a dedicated team to stay on top of the latest innovations in streaming bots and fraud prevention measures. This way, apps can be ahead of the game with updates and be aware of the latest app prevention software at an early stage.

Fraud can significantly put a dent in users’ trust levels, thus it is vital for businesses to ensure that they take actions to uphold this trust and keep the users informed of actions taken to combat bot fraud.

However, despite the damage fraudsters can inflict, bots can come and leave undetected. Bot detection requires a substantial amount of work, and in-house monitoring may take a significant amount of time away from IT teams.

For this reason, it is important for businesses to ensure that they equip their IT teams with adequate tools to detect and eliminate bots. By using machine learning and leveraging the complexity of anonymized sensor data from human-device interactions, bot detection solutions can behavioral patterns and distinguish between humans and bots.

From there, it will be easier to weed out the bots and gain control of the streaming economy.