Why Do You Like the Songs You Like? Startup Aims to Find Out

0

From left: Secret Chord Laboratories co-founders Scott Miles, David Rosen and Robert Abelow | Photo by LyVell Gipson

Scott Miles went to his first concert (Tesla and Poison) two months before his bar mitzvah.

“I was amazed at how everyone in the whole arena was sharing this same emotion,” the chief visionary officer and co-founder of Secret Chord Laboratories said. “And I just wanted to know how it’s possible how someone can move their fingers in a pattern and get this kind of emotional cohesion throughout thousands, tens of thousands, of people in this arena.”

This fascination with music and the brain sparked a career in research. In 2018, he joined David Rosen, a graduate of the applied cognitive and brain sciences program of Drexel University, to create the music analytics company Secret Chord. Rosen serves as CEO and Robert Abelow joined as chief operating officer in 2019.


All three of the founders are Jewish. Rosen and Abelow grew up playing in bands together in Cherry Hill, New Jersey. Miles and Rosen connected while conducting research at Drexel.

Secret Chord was one of 20 companies selected to participate in the Belgium-based Wallifornia MusicTech virtual accelerator program for music tech startups over the summer. During a series of workshops, industry leaders like Universal Music Group and Sony Music Entertainment judged the company’s product and strategy.

In August, the team beat 150 music tech startups from more than 30 countries to be named Wallifornia’s “Best Startup of 2020.”

The company’s artificial intelligence software, dopr, uses statistics and neuroscience to predict audience enjoyment of songs.

The premise started with graduate research at Drexel, where Miles and Rosen examined chords in 32 years of chart-topping songs and identified patterns of tension and release between verses and choruses. These are patterns of surprise, a level of unexpectedness that creates a sense of pleasure for the listener.

AI software could be fed songs to evaluate how newer patterns compared to those that came before them. Once the patterns of surprise were identified, they could be used to examine other songs and create statistical predictions about how they would perform among listeners. They could also be evaluated using metrics other than chords, such as harmony and lyrics.

Evaluating for patterns of surprise rather than patterns of similarity in successful songs allows the founders to find artists who are being innovative rather than creating more of the same.

“One of the things that is a common sentiment by people about pop music is that it gets stale, a lot of it sounds like sludge. It’s like the same thing over and over, and when a really great musical act or band comes through, it’s because they’re offering something that’s like new that nobody else can offer,” Rosen said.

Abelow said the beta version of dopr has proven to be an accurate predictor of song performance when tested by record companies and artists.

For the next phase of testing, Rosen said the company planned to use a National Science Foundation Small Business Innovation Research grant to test how closely their statistical calculations correspond with real reactions in people’s brains.

The ultimate goal is to help music executives understand what makes a song successful among target audiences. Surprise is beneficial in art, but that’s not always the case in business. That’s why producers under pressure to find the next big hit turn to artists who sound similar to recent chart toppers rather than seeking new talent.

By making it easier for producers to identify the qualities that make songs successful, the Secret Chord team hopes to lower barriers of entry to the music industry.

“We feel that in a lot of ways the music industry is still very top-heavy,” Miles said. “So we want to help kind of right that ship and make it so the people are able to have access to the artists who are doing the best stuff out there.”

The creators also intend dopr to predict preferences among different demographic groups determined by factors like age, gender and location. By using data from streaming platforms like Spotify and Apple Music, they will be able to recommend new music to listeners based on their preferences. This can also help new artists gain exposure by pushing listeners to explore the unknown.

“Something we’re seeing with a lot of streaming algorithms that are out there is they’re trying to keep customers on applications,” Abelow said. “So what ends up happening is you find the comfort zone for listeners, and then you serve them comfortable music that’s really safe and not taking a risk, not pushing their boundaries, because all you want to do is make sure they don’t hit skip,” Abelow said.

The company operates remotely, but Rosen has high hopes for future physical facilities.

“In my dream world, I see opening the national Secret Chord Laboratories, which is like the most state-of-the-art neuroscience lab slash recording studio in the world, in the Philadelphia area,” he said.

[email protected]; 215-832-0729

LEAVE A REPLY

Please enter your comment!
Please enter your name here