Picture of Daniel EK, Spotify’s Founding CEO
“Spotify Wrapped”, Spotify’s pivotal brand marketing campaign, has interestingly risen to a vertex case study regarding successful high-end marketing campaigns of the 21st Century. The rapid virality of the user-centric campaign, and how it has revolutionized the consumer-based marketing approach in the past eight years can be said to be a one-in-a-million occurrence.
Spotify Wrapped is easily one of the most exciting, to say the least, end-of-the-year reviews for over 640 million audio listeners around the World. The spikes in what makes the ingenious campaign a new media marketing enigma are streamlined in the personalized, rewarding user experience that it provides to its global user patronage.
However, 2024 has seen Spotify issue its most criticized year-in-review rollout since its debut oomph in 2017, when it accrued massive virality among millions of internet users for its colorful design cards and entrapping user metric display.
The early indications of Spotify’s in-house struggles beamed in this year’s lateness in releasing Wrapped. It took longer than usual, and enthusiasts of the year review campaign weren’t pleased with the delay (a grievance they made known through several queries on Spotify’s main social media accounts).
The giant streaming company takes great pride in its data engineering team’s ability to amass a crazy ton of user-streaming data upon click from the beginning of the year till the 31st of October each year.
Naturally, this time frame in collating and aggregating user data is supposed to provide all hands on deck with enough time to make the ‘Wrapped’ available before the holiday season and should be a walk in the park, not just because of the sheer size of processing power purchased to crunch these exabytes storage of data, but particularly because of the AI task force implemented.
However, unfortunately, the data collation and subsequent release to the public were far from prompt and accurate this year, and many users found Spotify’s algorithm wanting, with some saying it is inherently flawed.
Late last year, Spotify laid off 17% of its human task force leveraging on cost reductions, and adjustment to a slow growth pace. The layoff meant Spotify had to make up for workforce loss with an overt reliance on Artificial Intelligence.
For a global corporation such as Spotify, it is expected of the deployment of AI in its affairs. It’s non-negotiable as the World is fast moving in the direction of Artificial Intelligence. But what happens when Artificial Intelligence/ Machine models take the bulk load of Data collection, modeling, and crunching for final personalized deployment, as opposed to trained Data engineers and scientists? The result is this year’s poor, wrapped experience that has had many disappointed users threatening to boycott the streaming services if not improved upon.
In abrupt conclusion, Spotify Houses at least 100 million songs, 6 million podcasts, and 350, 000 audiobooks but the science behind Spotify data collection claims it can provide a breakdown of user data that reflects “Unique” consumer behavioral patterns through processed data collection, modeling, and crunching for final individual data deployment. But really, how plausible is this? It appears to this writer that this time, Spotify may have crunched more than its capabilities, and AI flops where humans trump.
