Kyle Kretschman is the Head of Economics for Spotify, a job he started during the pandemic after working as an economist with Amazon for several years. Before that, he was a doctoral student studying microeconomics at The University of Texas at Austin. One afternoon in October, we met over Zoom to discuss how he got from one point to the other and how he sees his liberal arts background affecting the work he does now.
What brought you to the economics program at UT Austin? What were you working on while a doctoral student in COLA?
I came to UT Austin because one of my economics mentors during my undergraduate years at the University of Pittsburgh had a connection to one of the faculty members in Austin and really recommended it as a Ph.D. program, not only for where it stood in the national rankings but also for how it treated its grad students. UT was really, really high on my list, and when I got accepted it was almost an immediate yes.
At the time, I thought I was going to become a macroeconomist studying labor market trends specifically around the European Union integration. But as I started to dive into more of the coursework, I realized that I liked microeconomics a lot more, so I moved from a macroeconomics focus to a microeconomics focus.
I love variety, so while I might be very specialized as an economist with a Ph.D., within that I like to poke around on things that I care about. For a second-year paper, I looked at the economics of education, and then I made another left turn into the economics of politics. I found a unique data set from the Federal Election Commission and ended up with a dissertation on how money turns into votes in congressional elections and how much candidates should spend. I liked that it hit some big structural economic issues, like substitution effects and optimal pricing, combined with a subject matter I cared about.
To go back a step, what originally drew you to economics, even before you began looking at UT Austin?
In high school I took the AP Economics classes, and all the topics intuitively made sense. The classes were easy because they were teaching me to put frameworks on how I already thought.
Then I went to the University of Pittsburgh as a computer science major, because I knew that would be applicable to a career and I had good mentors in the field. But after my freshman year, I just didn’t feel like it fit, so I transferred into a macroeconomics class taught by a former member of the White House’s Council of Economic Advisers in the Reagan administration, Steve Husted. He was able to marry the mental frameworks, the math, and also the stories of economics so that I was like, “This is it. I love this.”
I love that way of thinking about economics, as a set of tools and frameworks that you can use to tell a story. That feels like it gets at how economics fits into the liberal arts.
That’s exactly right. There’s also the mathematical foundation, and that’s the part that’s also problem solving. The math and the hard technical skills lead to what the job really is, and the job is storytelling. Economics is just different technical skills to tell stories.
Continuing this story: You finished your Ph.D. at UT Austin in 2011, and now you’re the head economist at Spotify. What’s in between those two points? How did you get from one to the other?
It was not a smooth, linear journey, even though I always knew I wanted to be a tech economist. One of my co-authors [in grad school], Nick Mastronardi, and I actually pitched a startup idea and he now runs that startup 15 years later. We also sent an email to the chief economist at Google, who was one of the few tech economists in the mid 2000s, and said “We will intern for you for free.” And he was like, well, actually I don’t really have that need this summer, but that shows this interest was always there.
When I left UT Austin, I had to decide between academia and the private sector, and the private sector won out. Part of the reason was that somebody with my variety of research ideas in grad school isn’t as attractive in academia, so I knew I didn’t fit the standard jobs. Instead, I took a job at a non-profit, federally funded research development center and did non-profit research on contract for the federal government for three years.
While I was doing that, I still had my eye on the tech sector. Then, economists in tech hit an inflection point where Amazon started hiring tech economists en masse. That was in 2014, and before I started at Amazon, I think there were probably about five Ph.D. economists who worked there. Now there are over 400, and I was part of a cohort of 10 or 15 that joined at the same time to prove that Ph.D. economists within tech can have major impacts.
Within Amazon, I spent six years generally working within the consumer finance organization, which meant analyzing and understanding the boxes and shipping things part of the business. Going back to your storytelling idea, one of my primary responsibilities was to figure out how New York Times headlines and major events impacted consumer demand. Whenever something external to Amazon happened — think of holidays and how people celebrate them — how did that affect consumer demand? My job was to build not only individual analyses but also software systems that can quantify the New York Times effect.
That was fascinating, I loved it. I didn’t honestly expect to leave Amazon, but then Spotify came with a great offer to work on something I care deeply about, which is the music industry and creators. They also asked, “Do you want to build again?” And I said, “Yeah, that’s a magic phrase for me, so yes.”
Now that you’re at Spotify, what are you working on?
For the past two and a half years we’ve been building the integration of Ph.D. economists directly within the business. They had, I think, maybe one Ph.D. economist in the whole company before I joined, but there wasn’t a discipline or a track for them.
I always say that Ph.D. economics and econometrics are really just flavors of data science. Then it’s a question of how we apply our statistical models specifically for the business-use case. It could be for marketing, which is the more traditional economist role of being more external facing, but that’s not what Ph.D. economists really do in tech anymore. We’re more heavily focused on internal strategy and pricing, and also how you can integrate statistical economic models into a product to make decisions.
One of the external marketing products we’ve been working on and that you can see is called Loud and Clear. It’s a website that explains how Spotify has accelerated the amount of revenue that’s gone into the overall music ecosystem. There are also internal product integrations, all in the vein of how we make the overall system more valuable to creators.
My job on a week-to-week basis is to make sure the people on my team and within my organization are solving the most high-value problems. So one of my key functions, I call it a translation function, is figuring out how we take economic research and econometric models and apply them in business.
I’ve built the Creator Economics team that focuses on the supply side of industry, and we’ve put another economist team in another business organization. Each economic squad is multidisciplinary with insights, data science, insights managers, business managers, and Ph.D. economists mixed in. It goes back to tying in to the liberal arts structure again: It’s not a siloed research function here, and it’s our job to understand the hypotheses and do the research process in a way that tells a story that can have a business impact.
You’ve already discussed this a bit, but how has your liberal arts education background shaped your work at Spotify? What is the overlap for you in your day-to-day or week-to-week?
The liberal arts and social science education that economists receive is what differentiates us in a tech company. It means we’re required to tell the stories, we’re required to think holistically, and we’re required to use critical thought.
UT Austin trained me very well to think about these things. Within the economics program specifically, I remember sitting in a class and one of the teachers said, “Can anybody actually explain how this would be applied?” Nobody could really answer that. And he goes, “Alright, we’re stopping everything,” and we focused on how the econometrics we were doing would be applied. That was a class on public finance with Dan Slesnick, and it was a real “aha” moment.
At Spotify, it’s the same thing: Figuring out how to take statistical models that help explain what’s happening within the ecosystem and within the company and figure out what the story really is so that we can then make positive change.
You say you’re working to make the overall system more valuable — meaning the music industry even separate from Spotify?
Both. Spotify works on what we call “stakeholder capitalism philosophy,” so we care about the overall stakeholders. It’s not just the company, but it’s also all the stakeholders who create value within it.
Because you can’t have Spotify if you don’t have musicians and labels making music to put on Spotify. You need creators that are thriving.
That’s right.
Were you musical before starting at Spotify?
I’m a huge music fan. I spent years in piano lessons, and then I had my mom take me out so I could play sports. It’s still one of the biggest regrets in my life, because I’m a huge music fan but not a creator or musician myself.
What advice do you have for students who are pursuing economics now or looking to join an economics Ph.D. program in the future?
I used to candidly tell people: be careful going to an Economics Ph.D. program with the final goal being going into tech. But that was probably around 2015, when there might have been hundreds of economics Ph.D.s total in tech. Now there are close to a thousand, and it’s actually a full career path to get a Ph.D. in economics and then go directly into tech. So, I’ve lightened that advice.
Generally, I think you should try to work backwards from where you want to go. If you want an economics term, call it “backward induction,” which is a nerdy way of saying to work backwards from your end goal. Did it work out for me in different ways? Yes, but I don’t think I was as strategic as I wanted to be, specifically during my Ph.D. program, where I really wasn’t sure what type of tech or non-tech or academic job I really wanted.
Think about when the big decision points in your life are going to be coming, and about where you want to end up, and then you can prepare.
Backward induction. I like the term, and I think that’s great advice.
I’ll also say, and this is unsolicited, but UT Austin is a great place to be trained to become a tech economist. In recent years, the department has structured not only a master’s program but also the Ph.D. program such that it can set you up for an academic job or a tech economist job. The proof points are places like Amazon, where I know that there are probably at least 10 Ph.D. economists from UT Austin. There’s very clear proof that this is a possible successful career path that UT Austin students can excel at.