Neuromarketing Principle: Online music popularity can be predicted from the similarity in brain activity of a relatively small sample size.
This guest blog was written by Nikki Leeuwis at Unravel Research and adapted for iMotions.com.
Would you have recognized the hit potential of Dua Lipa before it was cool? Or Drake?
Most probably, you won’t. There’s a lot of money going around in the music industry. But the question remains how much is invested efficiently, as this can only be addressed when the song has been released for a few weeks. It is estimated that only 10% of releases end up making a profit for the record label, with a financial loss incurring for up to 85% . Concurrently, the vast majority of music listening happens through streaming services such as Spotify, with revenue inherently tied to popularity measured through the number of streams, which of course can only be roughly estimated before release. But what if we had the power to predict what song will be in the charts for weeks, that everyone will be humming, and thus will be the blockbuster of the month and the cash cow of the record label?
Recent research by our client Unravel Research involving the iMotions platform has done exactly this. But first, let’s explain the theory behind the story.
Rewind To Ten Years Ago
Berns et al (2010)  had put some respondents in a fMRI brain scanner in order to measure the social influence on their preferences of relatively unknown songs. Two years later, some of the previously unknown songs were being played on the radio, while others were not. This triggered them to look back at the dataset in order to see whether they could have seen this coming from their initial measurements of the brain activity in the fMRI.
And they did! More specifically, how the respondents explicitly rated the songs didn’t have any predictive values in terms of popularity measured by sales data, but brain activity turned out to be an excellent predictor of the music’s popularity.
Now that it’s almost ten years later, there has been little continuation of this scientific forecasting of music popularity, which has been dubbed as neuroforecasting. On top of that, there hasn’t been a record label that uses these methods to make investment decisions based on neuroscience. Why is that?
Of course, a lot has changed since the research of Berns and Moore (2012) , especially since their methods for evaluating music popularity were based on the number of records sold. Nowadays, do you even remember the last time you went to buy a physical copy of music?
From Physical to Digital
The time gap between the seminal Berns and Moore study and today has essentially marked the rise of digital streaming platforms such as Spotify, whereby the music industry had to reinvent itself. Production and distribution costs diminished and with that, the battle for attention has only just started. Since digital streaming lowers the boundaries for distributing music, suddenly even smaller labels and lesser known artists can be explored without additional costs.
Simply put, the possibilities of digital streaming have broadened the musical interest of the public and the opportunities for emerging artists.
Can Neuro Predict Digital Music Streaming?
The power of neuromarketing in forecasting popularity in the entertainment industry has already been explained in this previous article in New Neuromarketing. Besides the previously discussed music study, there have been several studies that forecasted the popularity of movies based on neural activity during their trailers.
A frequently used metric within these predictive studies is the similarity of brain activity between respondents while watching the trailer. This metric is called neural synchrony and is related to – among others – box-office revenues and the success of package designs. Neural synchrony denotes the similarity of brain responses between multiple respondents, and it is detected through commonalities in the signal processing of brain signals measured using techniques such as electroencephalography (EEG), magnetoencephalography (MEG), and fMRI. It is seen in the literature to be indicative of successful communication of messaging or collective engagement with a stimulus, and thus predictive of popularity among a larger audience beyond data points from individuals.
The new study by Unravel Research has jumped into this cutting-edge metric of neuromarketing and entertainment, by applying the technique to what no one had done before: the predictive power of EEG-based neural similarity on music popularity on a digital streaming platform.
The study: measuring brain activity from newly released songs
The study consisted of using the iMotions software platform to expose 31 respondents to 24 fragments from newly released albums: the R&B album called “It Was Good Until It Wasn’t” by Kehlani and the pop album called “How I’m Feeling Now” by Charli XCX. The albums were chosen due to their convenient release date, therefore being unfamiliar to the respondents, and that they belonged to two different genres. The music fragments were subjectively sampled and contained the most distinctive part of the song, usually the chorus and/or the hook. Afterwards, respondents were asked if they had previously listened to the album – only one of whom had, so that respondent’s data was excluded from the analysis – and they were asked to rank the segments on a Likert scale. Then, the data was processed using iMotions with a custom-built R Notebook (which runs on the iMotions open-source infrastructure for signal processing algorithms) for brain activity analysis from EEG, and songs were monitored for the number of streams on the platform Spotify after both three weeks and ten months after the experiment.
What the researchers Leeuwis et al (2021)  found was that digital streaming popularity could be significantly predicted by neural similarity and its predictive value far exceeded subjective ratings of the respondents. Thus, the similarity in brain activity between respondents is highly predictable for group level popularity (40.4%). Especially when a song is released as a single, the predictive value is even higher (61.91%).
Figure 1. The distribution of singles across neural synchrony and plays (A) after either 3 weeks or (B) after 10 months log transformed. The blue line indicates the fitted regression line. The gray area indicates the confidence interval. The promotional single releases are indicated by blue triangles.
Why Can’t I…?
The fact that individual opinion is not as reliable for predicting population preferences has been already shown when Eminem won the 2017 EMA’s just because his performance was the last performance before the voting started. To overcome this and many other biases that underlie explicit survey answers, a more accurate measure was needed.
Neural synchrony provides just that. The study by Unravel Research shows how this metric might change the music industry by enabling efficient marketing budget distribution, promoting the right song as a single, and just hitting the charts right from week one.
The relevance of this research finding lies not only in the predictive power itself but essentially in the fact that previous research worked on fMRI measurements, which often costs up to six digits, while the price tag for EEG is affordable below $10,000. This broadens the audience that can access neuromarketing tools and thereby this methodology can be much more accessible for researchers to use.
Thus, with the validation of neural synchrony as a metric for popularity that works both with EEG and fMRI, the neuromarketing toolbox has been enriched once again.
Because neural synchrony is also being used as a predictor for the success of packaging and TV commercials, marketers might thereby benefit from this insight. Increase your marketing success by surpassing the method of explicit surveys and really focusing on the real decision center of the consumer: the brain.
As the brain activity within a small sample of respondents can be indicative of emotional experiences with a larger audience, this evenly holds for the population-wide product preferences of brands. If you are interested in conducting this type of research with EEG, you can read the full publication in Frontiers In Psychology, or request a live demo of the iMotions software.
 Vogel, H. L. (2020). Entertainment Industry Economics: A Guide for Financial Analysis. United Kingdom: Cambridge University Press.
 Berns, G. S., Capra, C. M., Moore, S., and Noussair, C. (2010). Neural mechanisms of the influence of popularity on adolescent ratings of music. NeuroImage 49, 2687–2696. doi: 10.1016/j.neuroimage.2009.10.070
 Berns, G. S., and Moore, S. E. (2012). A neural predictor of cultural popularity. J. Consum. Psychol. 22, 154–160. doi: 10.1016/j.jcps.2011.05.001
 Leeuwis N, Pistone D, Flick N and van Bommel T (2021) A Sound Prediction: EEG-Based Neural Synchrony Predicts Online Music Streams. Front. Psychol. 12:672980. doi: 10.3389/fpsyg.2021.672980