The secret weapon

How producing great music and producing great data science are eerily similar.

John Lennon

Snoop Dogg

Patti Smith

Marilyn Manson

What do a folk legend, a punk queen, a pioneering rap artist, and a shock rock legend all have in common? A secret weapon. A producer so legendary that he helped spur all of them to hit records: Jimmy Iovine.

What can an engineer, turned producer, turned manager, turned label owner, turned founder tell us about our approach to data science? Turns out, a lot. As data scientists, we need to bring multiple capabilities to our work, just like Iovine. Three particularly stand out: Understand the business context. Make the findings emotionally resonant. Find partnerships to keep our edge sharp. In this article I discuss how to accomplish these, taking inspiration from standouts in the music business. In this way I hope to show how we, too, have the capacity to be the secret weapon for our businesses to grow and develop.

Understand your artist

In music, artists are writing from their perspective, attempting to describe the world as they view it in their time. When Bruce Springsteen laments the struggles of blue-collar workers on America’s East Coast in the ’80s, you feel their desperation. When Tupac brings you into the poor neighborhoods of Los Angeles in the early ’90s, you understand the violence and strife in those communities.

Imagine being a producer, arriving at the studio knowing nothing about these circumstances — what can you do to prove to your artists that you understand their motivations and ambitions? Can you convince them to trust you so deeply that they believe their ambitions are yours too? Do your artists believe you have their back and are going to help them make the best records they can make?

Working with your business as a data scientist is no different. It too is existing in a context of competition, limitations, and goals. Deeply understanding this environment helps you understand the big picture. Knowing the framework that your business is operating in helps assure the work you do is relevant.

Have you ever completed an analysis or built a model only to have it underutilized? We all have. If you found yourself in this situation, perhaps your assumptions about the goals of your business weren’t quite correct.

One technique to help assure you’re in sync with your business is active empathy. Try on the point of view of your business by asking questions to find out the following: What will move them? What is their motivation?

If I’m able to interview stakeholders, one powerful question I often employ is: “What do you want?” or its variations, “What do you want most?”, “What are your near term (or long term) goals?” Simple questions like these can be a powerful way to cut to the center of what your business desires. From there, it is just a matter of tailoring your work. After asking these questions, ensure the analysis you choose to take on helps show a clear path to what your stakeholders are trying to achieve.

Make it raw

Early in Iovine’s career he worked with mainstream rock artists like U2, Tom Petty, and Stevie Nicks. In his work he helped push these artists to take their messages further, to a more visceral place. Later, as Iovine led the Interscope record label, he gravitated toward artists other labels were afraid of. These artists existed on the fringes, having even more shocking messages than anyone believed would be commercially viable. This led to the signing of powerful voices: No Doubt and Gwen Stephanie launched a new California sound, Nine Inch Nails popularized industrial rock, and Dr. Dre brought West coast rap to the mainstream.

Their messages and signals were clear, and they still stand today as canons of their genres. Iovine’s signing choices reflect an appreciation for a clear and pure message that drives emotion in the listener.

In data science, our messages can be complicated. Because of our training, we’re comfortable in a land of random forests, discounted gains, and standard deviations. In a world where we want to be fully understood, we often explain our methods and caveats before landing the punch line.

Doing this has a few consequences. First, it burns willpower. Your stakeholders have probably been in meetings all day, and now you’re asking them to understand deeply mathematical methodology. You’re asking them to concentrate to hear your message, which doesn’t help your cause. Second, they can only remember so much. What do you want them to leave the table with? Start there.

A key question I often use with my team is: “How should I feel about this?” and its variations “What does this mean to me?”, “What will this mean to our business?” and “Help me understand how to react to this.” My point in asking these questions is to help drive my team toward an emotional resonance with their stakeholders to prompt them to take action. As a data scientist, when you can help stakeholders know how to feel — whether it’s happy, concerned, or motivated, among others — they are more likely to invite you back to help them again, increasing your business impact.

Find your Dr. Dre

Jimmy Iovine is intensely passionate about sound. So much so, in fact, that many of the sounds we hear in popular music today are because of his obsessive fascination with getting the right mix in every song. When Dr. Dre produced the seminal rap album The Chronic in the early ’90s, it sounded so good to Iovine that he begged to know who engineered it. When Dr. Dre replied he did it himself, respect and friendship were born. That mutual love of sound and the changing music economy of the early 2000s led them to form Beats Electronics, a brand focused on preserving the intended sound of the music in the digital age.

By all accounts the relationship between Dr. Dre and Iovine is both respectful and challenging. They push each other to go further, change directions, adopt different tactics, and find new ways of exploring the music business they love.

In data science, the skills you know today, the techniques you’re employing, and the technologies you’re interacting with will eventually be outmoded. One thing that can keep you honest about your work and your capabilities is a collaborator who fiercely advocates for you — but also knows when to check you.

Data science doesn’t have to be an individual endeavor. Just as producing great music takes artists, engineers, producers, and audiences, producing great data science takes statisticians, ML modelers, data engineers, program managers, and stakeholders. It truly is a team sport.

Within our team here at Microsoft we spend multiple hours every week connecting on the work we do and offering feedback for how we can improve. We do this through organized sessions with leadership we call “Office Hours,” as well as contributor sessions we call “Peer Review.” We also encourage our team members to mentor and mentee each other to keep ideas and knowledge flowing. We think about these interactions as “jam sessions” where we can bring our early tracks and refine them. When person-to-person contacts aren’t available, we leverage our knowledge catalogs to reference previous work.

Great songs and great analytics don’t exist in a vacuum. Start building your team of visionary collaborators today and see how much further your ideas can go.

Release the album

Music and numbers have so much in common. Some may even argue that music is numbers — frequencies that hit our ears at differing vibrations. These waves result in sounds that can transport us to a new place, change our mood, or give us a new way to think and feel about the world around us.

Data science has this same transformative power. Produce your work in a comprehensive context. Share your findings via clear and evocative signals. Improve your output through deep and challenging partnerships. Be the secret weapon who makes noise with impact in our industries, communities, and beyond.

Natalie Duryea is on LinkedIn.