How Spotify Changes Your Music Preferences
I spoke to the man behind the Spotify algorithm

Hi,

This past weekend, I finally got around to reading a book I had been eagerly anticipating for months: You Have Not Yet Heard Your Favourite Song by Glenn McDonald.

McDonald is arguably one of the most intriguing figures in the music industry.

Why?

Glenn McDonald is the former 'Data Alchemist' at Spotify and the creator of Every Noise at Once, a project that literally maps music genres globally, provides critical input for the Spotify algorithm, and helps listeners discover new music.

Today, I bring you an exclusive as I asked him several questions (plus a brief preview of this new book).

📈 7 Questions for the Former Data Alchemist of Spotify

Can you tell me more about your role as 'Data Alchemist' at Spotify and how it came about?

I was part of a startup called The Echo Nest that Spotify eventually acquired. Startups often have quirky roles because you're usually making things up on the fly without really having the right resources. I started calling myself a Data Alchemist because if a Data Scientist was looking for the truth, an Alchemist was just trying to transform things into other things. That was a much better description of what I was doing. And then after we got acquired, I just stubbornly kept doing it, and it took a really long time before anybody stopped me.

How do you see the future of algorithmic recommendations in music? What improvements or changes do you anticipate?

I don't predict; I hope. What I hope happens with music algorithms is less "recommendation," less passive listening, and more active curiosity and exploration facilitated by algorithmic tools under more listener control.

What do you consider the biggest advantages and disadvantages of the current streaming model for artists?

The biggest advantage is that every artist is just a few clicks away from any listener. It’s now possible for any artist to reach any audience in a way that was never possible in the physical era. The disadvantage is that this tantalising possibility has attracted many artists, but statistically, very few of them actually become huge. And the tools, labels, and streaming services are still mostly focused on the stars who do become massive, rather than helping to build the communities that would allow smaller artists to have viable careers without needing to be gigantic.

How is AI affecting the way music is created and consumed? Are there specific AI applications that give you hope?

Large Language Models are astonishing and fascinating but terrible at many of the things we're currently trying to do with them. The raw potential seems enormous, but we have a lot of work to do to figure out how to harness it productively and reliably. (Which is basically the same thing I say about streaming music in the book!)

What advice would you give to artists trying to break through in the current digital age?

Build communities. Find communities, join them, help them thrive and grow so that they can help sustain you. Communities can be physical, virtual, geographic, or philosophical—people find many ways to organise themselves. But almost nobody makes it on their own, and it’s lonely even if you somehow manage it. Connect with other artists, fans, curators—everyone. Music has always been a social art, and nothing “digital” can change that.

Can you share an example of a musical community you discovered through Every Noise at Once that you found particularly fascinating?

I've become an obsessive fan of wedding and party music from Limpopo, a province in the northeastern corner of South Africa. I’ve never been anywhere near there, but I stumbled upon this music because it kept appearing around the edges of Maskandi, a form of Zulu folk-pop that I already knew and liked. I could tell it wasn't Maskandi, but I didn’t know what it was, and nobody could tell me. It took me a while to figure out the story, but once I understood, of course, I added it to the everynoise.com map. It's called "sepedi pop," after the ethnic/language group whose weddings and parties they are.

Spotify offers podcasts and audiobooks alongside music. Do you think this diversification is necessary for a profitable model?

Spotify's profitability was thankfully never my responsibility. As a music person, I would have preferred to focus solely on creating a truly great music service and let podcasts and audiobooks thrive on their own platforms, run by people who love those things the way I love music.

🧠 How Streaming is Changing Music
Some (other) interesting topics he discusses in the book:

• What the streaming service knows about you: Spotify collects extensive data about your listening habits, including demographic data and location, to make recommendations. He explains this in detail in the book.

• 'Cheating' the algorithm: Fans and artists sometimes try to influence or manipulate the algorithm (Spoiler: the actual benefits are often limited).

• Cultural differences: From the secrets of "russelåter" in Norway to Christmas in the Philippines, he describes how music consumption differs worldwide.

• Fraud in streaming: Despite technological advancements, fraud remains a problem in the music industry, albeit in new forms.

• Trends: Jazz and ASMR: McDonald explores whether traditional genres like jazz are actually disappearing and how new trends like ASMR are emerging.

Critical Questions and Future Vision

Another important point McDonald raises is the financial aspect of streaming. He discusses how the current payment system works and its impact on artists, advocating for a pro-rata system based on listening time rather than the number of plays. He believes this system would be fairer for less popular artists.

McDonald also emphasises the importance of categorising music. His project, Every Noise at Once, emerged from a need for a genre system and grew into a way to discover and literally map musical communities and cultures worldwide.

He also delves into the future of the music industry and the role of AI in music creation. He expresses concerns about AI-generated music but also highlights the potential that new technologies offer to stimulate creativity.

According to McDonald, algorithms should foster listeners' curiosity and bring communities together.

In the coming weeks, I will try to share more interesting themes and insights from the book.

But, a tip in advance: just go and read it yourself.

The book is a must-read for anyone in the media, entertainment, and music industry.

Check out the book here

📌 Reading, Listening, and Viewing Tips
Universal, Sony, and Warner are suing AI music services Suno and Udio for copyright infringement. Why? Suno and Udio may have used copyrighted material without planning to obtain licences or provide compensation to rights holders.

Listening Tip: The brand-new podcast by music journalist Mark Sutherland: The Money Trench - The Music Industry Podcast. The first episode features an intriguing guest: Lorna Clarke, Music Director at the BBC, who gives a behind-the-scenes look at Glastonbury. What does it take to broadcast such a mega event? Listen on Spotify here.

Job Opportunity: The Chassé Theater is looking for an Online Marketing Strategist (32-36 hours).

Check out the Every Noise at Once map with all the genres Glenn McDonald talks about in his book: https://everynoise.com/

Listening Tip II: Want to know more about the book and Glenn McDonald? He was a guest on The Gist, the daily American news podcast by Mike Pesca. Listen to the episode here (starting at minute 6).

Vind je deze nieuwsbrief interessant?

🔗 Voeg me gerust toe op LinkedIn voor meer nieuws, artikelen en tips.

📖 Vind hier alle artikelen over streaming, podcasting en marketing.

📩 Wil je reageren, heb je een tip of een vraag? Stuur me gerust een berichtje d.m.v. een reactie op deze mail.

These cannabis gummies keep selling out in 2023

• Mood is an online dispensary that has invented a “joint within a gummy” that’s extremely potent yet federally-legal.

• Just 1 of their rapid onset THC gummies can get you feeling right within 5 minutes!

Order Now