How AI & Data Analytics could have prevented an amazing Spotify fraud

We`re not sure if you already have heard about this and if not, keep in mind that those couple of minutes will be more than worthy.

The Spotify Fraud

Unknown scammer, allegedly from Bulgaria generated about $3m revenue out of creating and continuously playing a couple of playlists with tracks with an average length of 43 seconds. Spotify is paying artists about $0.004 per play and the fraudster has registered around 1200 premium Spotify accounts, continuously rolling the playlists.

It is believed that such ‘performance’ was achieved via bots automating the skip and play game regarding the Spotify policy to pay for a listened track only its 30 seconds or more. This way, the playlists called ‘Soulful Music’ and ‘Music From The Heart’ became played so many times they made it to number 84 globally and 22 in the US in the playlists charts.

How was the fraud spotted?

You might guess some sophisticated AI or analytics algorithm played a huge role in spotting such abnormal behavior. Not at all! It was the unbiased human expertise displayed by some major music labels insiders. Music Business Worldwide journalist Tim Ingham burst the story and pointed out that in September 2017, the streaming service has deleted the playlists.

Each week, Spotify is sending tracks and playlists performance figures to industry insiders like major labels. By climbing high at playlists` charts, they caught the label`s attention. Surprisingly, the songs weren’t so exciting or addictive and their copyright was owned by a single person. In his feature, Ingham quoted different music industry sources stating that such unusual tracks, playlists, and behavior are nothing less than a scam, which took at least a $1m. out of the Spotify` royalties pool.

How could AI and Data Analytics help in such scenario?

Analytics and its tools are just perfect for spotting abnormalities and deviations in large data sets. If Spotify has set specific data alerts on unusual behavior, the fraud would be spotted far earlier than September 2017.

We already discussed how savvy data scientists and engineers helped Spotify to become a market leader and by utilizing the tools of data analytics, machine learning and AI they can easily prevent such fraud of happening again.

Well, it happened. How to prevent it?

As it already happened, we are sure that the Spotify` data experts will be able to prevent such fraud to happen again easily by carefully monitoring the length of the songs, the playlists they form and the accounts royalties progress on a weekly or monthly basis. As we are not insiders at the streaming service data infrastructure, we are sure a lot more can be done.

One more thing, we are curious what the YouTube execs are thinking of this story as they also pay royalties to their content creators.

Leave a Reply

Your email address will not be published. Required fields are marked *