Results Of Our Three-Month Pre-Midem Prediction Test
Three months ago, we took 5 newly-released songs and used our music-focused algorithm to predict how many streams each would amass by the start of Midem in the first week of June. It's now been 3 months since we made our predictions — how do you think we did?
The idea for this test came to us when we applied to present Hyperlive at the Midemlab pitching competition (check it out, there are some great new companies pitching this year). As part of our preparations, we decided to select a handful of songs released in the same week and see how accurately we could forecast their performance by the start of the competition.
We selected songs in a range of styles and by artists at different stages of their careers, including Raw by the hotly-tipped Sigrid, Can't Deny Me by rock legends Pearl Jam, Alone by Halsey, Done For Me by Charlie puth, and YES by X-Factor UK winner Louisa. All of these tracks were released in week 11. To give you an idea of the task at hand, here's how the situation looked back in March when these songs were released (our predictions are in yellow, total streams at the time — where visible — are in white):
As you can see, previous performance was virtually non-existent, and certainly unlikely to be a good predictor of future performance, mirroring the typical situation on the day a new song is released.
So we analysed the music, made our predictions, and sat back. As usual, we predicted the combined total of Spotify and YouTube streams each track would amass, with the twist on this occasion being that we set a specific time-limit to those predictions, i.e, how many streams they'd amass by today 5th June 2018. And remember, we made these predictions back in week 11, the same week all of these songs were released.
Although we didn't make the cut for the competition (boo!), we thought we'd share our predictions anyway since we're pretty proud of how they turned out.
So, that was the situation at the time of release; here's the situation today.
Just to be clear, all predictions were based on an analysis of the musical content of each track, and, since they were made 3 months ago, obviously with no prior knowledge of actual performance.
So what do you think? If you're a fan of any of these tracks, how did our predictions compare to your expectations?
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.