Kevin Novak joined Uber in 2011 as employee number 21 and engineer number seven, and by 2014 he was the company’s head of data science. He speaks proudly of that time, but like all good things, it had come to an end, and by the end of 2017, after the company had done what he wanted, he was gone.
Initially, he picked up the pace of angel investing, something he had already focused on on weekends and evenings, eventually building a portfolio of more than 50 startups (including financial technology company Pipe and self-checkout company Standard Cognition).
He has also started advising start-ups and venture firms — including Playground Global, Costanoa Ventures, Renegade Partners and Data Collective– and eventually realizing that he wanted to spend the rest of his career as a start-up investor, Novak decided this year to launch his own venture capital firm in Menlo Park, Calif. Rackhouse Ventures. In fact, Rackhouse just closed its first fund with $15 million, led by Curtis Chambers, Uber’s first head of engineering, Steve Gilula, former chairman of Searchlight Pictures, and Cendana Capital of the fund. Many of the venture capitalists Mr Novak knows are also investors in the fund.
We caught up with Novak last weekend to talk about the new car. We also talked about his tenure at Uber, where, mind you, he was instrumental in creating surge pricing (though he prefers the term “dynamic pricing”), and you can hear a fuller discussion, or check out the excerpt below, which has been lightly edited for length and clarity.
TC: You planned to be a nuclear physicist. How did you end up at Uber?
KN: At undergraduate level, I studied physics, mathematics and computer science, and when I got to graduate school, I really wanted to teach. But I also love programming very much, and I love applying physics concepts to programming, and the nuclear department has the biggest supercomputer time allotment, so that ended up driving a lot of my research — just having the opportunity to play with computers while doing physics. So [I] studied to be a nuclear physicist and was funded very indirectly through the research that eventually became the Higgs boson. With the discovery of the Higgs, this is very good for humanity and absolutely terrible for my research budget… .
A friend of mine heard about what I was doing and knew my skill set, and he said, “Hey, you should see this Uber taxi company, it’s like a limousine company with an app. There’s a very interesting data problem and a very interesting math problem.” So I ended up applying, [even though I made the big no-no on startup applications], wearing a suit and tie to the interview.
TC: You’re from Michigan. I grew up in the Midwest, too, so I understand why you’d think people would wear suits to job interviews.
I got off the elevator and the friend who encouraged me to apply said, “What are you wearing?” But I was invited in, became a computational algorithm engineer — a title that preceded data science trends — and spent the next few years living in engineering and product, building data functions and… Like our ETA engine, which basically predicts how long it will take Uber to get to you. My first project was working on tolls and tunnels, because figuring out which tunnels Uber went through and how to establish time and distance was a common failure. So I spent three days, with a bunch of cell phones, driving from Big Dig in Boston to Somerville and back to Logan, collecting GPS data.
I know a lot of very random facts about Uber Cities, but my big name is dynamic pricing. . It proved to be a very successful cornerstone, ensuring Uber’s availability.
TC: How does that work when you tell people you invented dynamic pricing?
KN: It’s a very quick litmus test for the underlying enthusiasm for behavioural economics and finance. The Wall Street crowd is like, ‘Oh my God, that’s so cool,’ and a lot of people are like, ‘Oh, thank you, yeah, thank you so much, great, you buy a round of drinks’ and stuff like that. . [laughs].
But data also became the incubator for a lot of the early special projects, like Uber Pool and a lot of the ideas around, well, how are you going to build a scheduling model so that different people have a set of ride requests? How do you batch them together effectively in space and time so that we can get the matching rate right, [so this] project is profitable? We did a lot of work looking at the theory behind the Uber Eats delivery model and thinking about how we could apply our experience in shared mobility to food. So I looked at these products from the point of view of the first three people scribbling on notepads or thinking on laptops at lunch [and these products] ended up becoming these national conglomerates.
TC: You worked for Uber Freight for the last nine months of your employment at Uber, so you and Anthony LevandowskiThis businessIs booming.
I joined a company of 20 people and we were approaching 20,000…… I kind of missed the energy of small teams and felt like I had reached a natural stopping point. And then Uber 2017 happened, and there was Anthony, there was Susan Fowler, and Travis had this terrible accident in his personal life, and his head was clearly not in the game. But I don’t want to be the worst quarter in the history of the company is famous for its move in, so I finally took a year, basically maintained the unity of the band, and trying to find out what I can do to keep me is any small business integrity and incentives, as well as sympathy and all kinds of good sense.
TC: You left at the end of that year and you seem to have been busy ever since, including now, launching this new fund with outside support. Why is it called Rackhouse? I see you use the brand Jigsaw Venture Capital when you invest your own money.
Yes, even a year ago, I had set up an LLC, and I was “marking” my portfolio to market, sending quarterly updates to myself, my accountant and my wife. This is one of my exercises in coaching managers, and I believe that if you can develop a few skills at a time, you will grow most effectively and successfully. So I’m trying to figure out what it takes to run my own back end, even if it’s just moving my money from my checking account to my “investment account” and writing my own portfolio updates.
I was also very excited about the possibility of launching my first external fund under Jigsaw with other people’s money, but there was actually a fund [called Jigsaw] in the UK when I started talking to LPs and said ‘look, I want to do this data. Look, I want to do this data fund, and I want it to be early stages, “I got a call from them,” and we just saw Jigsaw do this Series D round at Crowdstrike. I realized THAT I would be competing with other Jigsaw companies from an idea-share standpoint, so I thought I’d create my own unique brand before things got too big and too crazy.
TC: Have you included your angel-backed trades in the new fund? I see Rackhouse has 13 portfolio companies.
There are a few that I have agreed to move forward and build positions for the fund, and we are just working on the technicalities of that right now.
TC: The focus is on machine learning and artificial intelligence.
Yes, I think there are amazing opportunities outside the traditional focus areas of the industry, and as long as you can find rigorous applications like ARTIFICIAL intelligence, the competition will also be greatly reduced. Those [deals] that are not in the range of many [risk] firms are the games I want to play. I think the opportunity — no matter what industry, no matter what region — is biased in favor of domain experts.
TC: I wonder if that also explains the size of your fund — you want to stay out of the hit zone of most venture firms.
I wanted to make sure I was building a fund that would allow me to be actively involved in the company at an early stage.
Matt Oko and Zach [Data Collective] are good friends of mine — in fact, they’re mentors and the little LP of the fund, and they talked to me about how they got started. But now they’re managing over a billion dollars in assets, and the people I [like to back] are two people who are working part-time and ready to take risks, and [a company the size of Data Collective] has basically priced itself out of the startup and pre-seed phase, which I like. At this stage, I have a lot of useful experience. I also think at this stage, if you have domain expertise, you don’t need five quarters of financials to be trusted.