Predicting The Sound Of Summer 2019

Hyperlive Sound Of Summer 2019 Prediction Accuracy.jpeg


As the summer winds down and we approach the cooler months of Autumn, we thought we’d look at how accurately our algorithm predicted the soundtrack to three of the hottest months on record.

To heat things up a little further, instead of predicting performance after the same amount of time has passed for each track (such as, say, one month after release), we thought we’d try predicting performance on a particular target date, i.e., the same date for all tracks, regardless of when they were released.

Our test tracks were drawn from the playlist Summer Hits — The soundtrack of the summer 2019 by Spotify. This playlist, published on 30th June 2019, contains 55 tracks presumably envisaged by Spotify to capture, to a greater or lesser extent, the sound of summer 2019. These tracks were released between 7 and 70 weeks ago, and represent a range of artists, styles and genres.

We decided to predict how well each of them would perform by a specific target date — today, 19th August 2019 — and compare our predictions with their actual performance.


We analysed each track’s musical content, passed it through our music-driven algorithm, and predicted the total number of streams it was likely to have amassed on Spotify by today.

Comparison of our predictions with actual performance revealed that, for 71% of tracks, prediction error was less than 10% of artists’ typical performance range. Across all tracks, median artist-standardised prediction error was just 5%.

In other words, our predictions closely matched actual performance.

Next, based on the number of streams we predicted each track would have amassed by today, we classified it as being a Minor Hit, a Major Hit, or a Cultural Phenomenon. We then did the same for the actual number of streams amassed, and compared our predictions with actual performance.

Overall, we correctly classified 82% of tracks. Digging a little deeper, we correctly identified 90% of Minor Hits, 87% of Major Hits, and 70% of Cultural Phenomena.

What's more, we correctly identified 7 of the 10 most commercially successful tracks on the playlist, including the surprise hit Sucker by the Jonas Brothers, Ed Sheeran's top track, I Don't Care, and the biggest hit on the whole playlist, the Billion-streamed Better Now, by Post Malone.

At the other end of the spectrum, we correctly identified all 10 of the 10 least commercially successful tracks.

Finally, we correctly predicted that the most recently-released track on the playlist, Ed Sheeran's Beautiful People, would become an official Cultural Phenomenon by the target date, despite having been released just 7 weeks prior.


In sum, we were able to predict the sound of Summer 2019 with a high degree of accuracy.

In so doing, we demonstrated that we can not only accurately predict commercial performance of new music in the short- to medium-term, but over significantly longer periods, too.

What’s more, we can accurately predict how commercially successful a track is likely to be by a specific date, offering added value to artists, labels and rights-holders wishing to quantify the hit (and earning) potential of their music.

If you have any questions or comments about these results, or would like to learn more about how we calculated them, we'd love to hear from you

For more general info about how Hyperlive quantifies listener engagement and forecasts future performance, check out our Home or About pages.

About Hyperlive. Music isn’t something we just listen to — music is something we experience with our heart and soul. Hyperlive captures this experience by modelling a range of neurobiobehavioural responses to music as well as the psychological processes that underpin them. This gives us a deep understanding of what drives musical engagement on a fundamental level, allowing us to quantify, model and predict that engagement — and the musical features that motivate it — with unmatched levels of precision.