Q&A: Wyatt Marshall
Q&A: Wyatt Marshall

Q&A: Wyatt Marshall

The SoundCloud data wrangler talks about the world of information in the numbers.

When you think "MBA" you might picture a highly analytical numbers person who loves to examine the gears of a business. When you think "journalism degree" you might picture an intrepid fact-finder who relentlessly turns over stones. When you think "metalhead" … it's possible you're now picturing Beavis and Butthead. 

Or you could stop exhausting your imagination and have a conversation with Wyatt Marshall. Since 2018 he's occupied a number of positions at SoundCloud—nearly all of which had the word "data" in the title—and he has now been the director of music intelligence and analytics there for two years. And it doesn't take much imagination to instantly see how those three elements of his qualifications make perfect sense when blended for the purpose of his role. He'd be no good at his job if he didn't love digging into data and spreadsheets, and he wouldn't know what to do with the information he extracts from the data if he didn't have a journalistic storytelling instinct. The metalhead part, well, Wyatt doesn't resemble Beavis or Butthead in the slightest. But his passion for music is what enables him to connect with the listeners and creators who use SoundCloud. 

Here, he tells us about the ways data is essential to digital music, but only when used in combination with human intelligence.  

Tell us a little bit about your job—how do you describe it to someone who's not involved in the back room of digital music? 

At a high level, I generally say I try to keep the pulse of the SoundCloud platform. That includes making sense of what’s happening musically from an editorial, storytelling angle of the kind that makes for good music journalism. 

It helps to underscore that SoundCloud is more than just a place to stream music. It’s a place where listeners come to find their new favorite artist before they’ve broken into the mainstream, and it’s a place where budding artists can find inspiration, upload their first track and have it exist alongside tracks from the biggest names in music. In many cases, fans are artists and artists are fans, and they express their feelings toward the music they discover with the social features built into SoundCloud—likes, reposts, shares, and so on, which produces data.

So as a two-sided platform that’s fueled by both artists and fans, we want to understand what’s resonating with listeners, and how artists are using the platform and finding inspiration from one another and how they’re building communities. As the director of music intelligence and analytics, I try to make sure we’re keeping up with all the exciting things happening on the platform.

Data analytics around streaming music is still a relatively new science, since streaming music has barely passed its 20-year birthday. How is it evolving? 

Because it’s relatively new, there’s a lot of room for novel thinking to take place. Watching the interplay between understanding what’s resonating with listeners and what’s inspiring artists has become increasingly important. I’m sure as the landscape continues to shift, with new modes of music-sharing and creation, we will all have to continue to stay on our toes.

I think we’ll see more emphasis on understanding fandom and identifying who is going to take the leap from casual listening to more tangible support. Data-driven A&R research is going to get stronger and stronger. But it isn’t all just numbers, and those who will do the best here, I think, will be those who bring a true appreciation of music and musical and cultural context and expertise to bear. 

What are the most important questions that data can answer for a music service?

Data can help you understand what’s striking a chord musically with different listeners in different places, it can tell you what kinds of features help turn people on to new artists, and a lot more. We’ve all seen the discourse on the economics of streaming, so I think we all realize how important it is to understand the difference between casual listening and fandom, and what signals indicate real fandom. That understanding uncovers deeper connections between fans and artists, and makes for a richer and more rewarding experience for both.

What's the coolest part of your job? 

What continues to fascinate me is the depth of discovery, and how small pockets of artists experimenting with diverse sounds can cast aside preconceived notions of genre and song structure and build fanbases that grow over time. Next thing you know, a small community that caught your attention for one reason or another is suddenly the next big thing.

As you go exploring, you discover incredible music, things you never knew existed. As a listener, you can chart a path from dark trap to Vinahouse and hit about 50 other totally different sounds along the way. Many artists who are uploading boundary-pushing music are forming communities with like-minded artists on the platform. Another thing that’s been cool to observe is how certain sounds have traversed international borders. Sounds can take hold in one place and subsequently catch on with artists who reside in another country, and you can sort of watch the spread of highly specific sounds and styles in real time.

What are some of the limitations of data?

While data signals can point us toward tracks or artists that seem to be seeing greater activity than others, and maybe other data can give an indication of quality or sonic characteristics, there’s no accounting for taste. Music hits you and me in different ways, and it takes human ears and brains to pick up on the emotional connection that draws us to music in the first place. I don’t think data can predict the impact that music will have on the individual, especially when it comes to exposure to different sounds that are novel to a listener. I’m a metalhead, and an algorithm based on my listening habits as a middle schooler couldn’t have predicted that a chance hearing of Amorphis would have led to the lifelong connection it has.

What do you think will be the impact of AI on the field of streaming music, and of playlisting and music creation?

I’m no expert here, but I personally believe it’s going to play an increasingly huge role both in trend and predictive analysis, and we’ve already seen a great deal of coverage about AI in music creation, playlisting, and more. With audio fingerprinting AI can give curators the tools to more effectively sift through millions of tracks to find the songs that most perfectly fit a given sonic palette or mood. In terms of personalized algorithmic playlisting, we’re going to see playlists that pick up on sonic traits and moods that fit our listening habits with greater and greater uncanny nuance. 

Your education includes a masters in journalism and an MBA. Was applying those skills to music always a goal? How do they intersect? 

I was writing about music and a variety of other topics as a journalist before I went to business school, and when I was putting together my application essays I wrote about how I wanted to connect would-be-fans with cool, underground music that wasn’t served well by streaming platforms. That was rooted in my work as a journalist and as a fan. At SoundCloud I quickly became aware of the need to see things from a big data point of view, though, and I found that there was room to plant guideposts rooted in music expertise into the data to help better make sense of it all. 

The need for music contextualization applies to both someone analyzing the back end of a DSP and to a listener navigating an app. I think cultural context and storytelling really helps bring music alive, and helps move someone from being a casual, passive listener to becoming a fan of an artist or genre. So I try to sit at that intersection of analytics and editorial contextualization as often as possible.

How do you, for your own personal purposes, discover new music? 

When I’m not in my SoundCloud chair, I’m far, far from scientific in my discovery methods. I write a metal column, the Black Market, over at Stereogum where Ian Chainey and I highlight our 10 favorite metal tracks every month, so I’m constantly looking for new metal. For that, I’ll dig around on Bandcamp, where metal has a really strong community, there are some YouTube channels that are really useful sources, and I’ve got a stack of promos landing in my email every day that I could never hope to keep up with. It may sound a bit basic, but a good band name will catch my attention, and so will album art—metal has such a strong visual language, and if a band nails the look and vibe, they’re onto something, so I’ll give it a listen and am usually not disappointed. It’s hard to beat the feeling of stumbling across a new gem that clicks within seconds and will stick with you for months and years.

More Blog Posts

Dave Lankford on Streaming Innovation

Dave Lankford on Streaming Innovation

Rachel Saywitz
The former VP of Advanced Innovation at Disney chats with us about growth in the streaming video space and the everlasting power of a good story.
Downtown Music’s Molly Neuman on the Music Industry’s Biggest Transformations

Downtown Music’s Molly Neuman on the Music Industry’s Biggest Transformations

Hannah Elliott
The CMO of the independent music publishing company reflects on her own career as an advocate for indie artists.
The Pleasure of Deadlines According to Annie Zaleski

The Pleasure of Deadlines According to Annie Zaleski

Alison Aves
Flexibility is not your friend when you're juggling this many things.

Let's Talk

Do you have a project that would benefit from a world-class team of data analysts, pop culture writers, and marketing strategists? We’d love to hear from you.

Get in Touch