Zach Weinberg is a Co-Founder of Curie Bio and Operator Partners. Zach was previously Co-Founder, President, and COO of both Flatiron Health (acquired by Roche in 2018 for $2B) and Invite Media (acquired by Google in 2010 for $81M). Zach is one of the most successful entrepreneurs and investors in healthcare technology and we are delighted to share this interview with you.
Tell us about Flatiron Health and your original vision for the company?
The theory behind Flatiron is that there is tremendous hidden value in the electronic health record (EHR) for oncology patients. In particular, the clinical data that exists inside of the health records – think physician notes, radiology reports, pathology summaries, and outcome data.
If the EHR data is aggregated at scale and cleaned appropriately, it can be used for a variety of really interesting things – everything from clinical trial design, to post-approval evidence generation, to hypothesis generation.
It was very difficult to do that work, both to get access to the data and to clean the data in a consistent, high-quality way at scale. The thesis was about pulling that off in an efficient manner by building interesting, important tools for the cancer centers and building robust technical infrastructure to deal with the data curation problem. We then built a pharma biotech business on the other side to monetize the insights.
What were some of the biggest challenges you faced while you were building the company?
It was everything from:
- How do you design a sales process that resonates with independent cancer practices? How does that differ when you take it to an academic medical center and hospital system?
- What is the variability and nuance in the data?
- How do you articulate the right use cases for Pharma that matter? How do you think about pricing for Pharma?
- How do you do all of this while building an organization that’s half medical, half technical, while they’re running towards the same thing?
How to do all that without going bankrupt? Because it’s very expensive to do.
It was managing twenty parallel moving parts. Had we known how complicated it would have been, I might have been scared away. We were somewhat naive in doing this work. We thought it was an important problem and we thought we had an interesting insight that people had missed, which had to do with the data itself, how to use the data. We were pretty sure people had missed it, which turned out to be true.
Building connectivity to the cancer centers must have been challenging, and many other companies have tried. What was your approach to making that happen?
The biggest thing we did is we really understood each stakeholder. This is what this (person, group, or entity) cares about, and here’s what’s important to them. Let’s make sure we build the right thing for them rather than what we think is right for them.
We had a deep understanding of what it meant to be a community oncology practice, what it meant to be an academic medical center, and what it meant to be a pharmaceutical company that had an on-market therapy versus a pipeline therapy. A new launch versus one that had been on market for quite some time. We got very good at understanding the customer, given the number of different customer phenotypes. It wasn’t like a software company, where we only sold to the IT team. There were layers of different customer types, and each one had nuanced incentive structures and sophistication levels. It took some time and we were really in the weeds.
What did you learn about enabling successful collaboration between the software and medical teams when building Flatiron?
It comes down to leadership and culture. It’s not an org design problem, it’s a hiring problem, then a culture problem in terms of how you guide people to what’s acceptable behavior. We struggled with this, but the advice I give to all the founders that I work with is that you need product and engineering leadership that have an appreciation and a passion for learning about medicine or healthcare, which is complicated. Then you need to invert into the other side of that, you need the medical or healthcare leadership who don’t look at technology as this black box, but wants to understand it. If the leaders want to understand, the team members will follow suit.
Founders have to be very clear about it – we’re not just a healthcare company, we’re not just a technology company, we are both, and it’s important for us to be both. We used to say this all the time, we have to be good at both of these things for this to work. So one is not more important than the other they both have to exist. It’s hard. It’s part of why you don’t see major healthcare technology company outcomes.
Do have any observations on how the Real World Evidence (RWE) and Real World Data (RWD) space have evolved since you first founded Flatiron?
Well, for starters, it has a name. It was not a thing when we started. People called it retrospective data or observational data. This whole Real World Evidence branding kind of evolved. I’m not really 100% sure how.
I think it’s probably in the hype cycle, versus where it will end up. As with all new things, people overestimate what’s possible in the short run, and they underestimate it in the long run. I think RWE has the ability to make certain decisions better or faster or cheaper. In particular, in drug development, but there are limits to what’s possible because the data is fundamentally unrandomized. So there are only so many questions that matter that you can answer in a sophisticated way. There’s enough of them, they’re interesting, but at a certain point, you just run out of space.
Part of that is that Pharma is relatively consolidated. How many on-market therapies are there or will there be? That’s your market right there, and that was always the biggest frustration to me about Flatiron. Had there been 10x, the number of customers, we would have built the exact same thing. What we built for the top 13 or 14 pharma companies wouldn’t have been different if there were 150 top pharma companies. If there were 150 customers, we would have had a much, much bigger business. But there weren’t, and there aren’t. So I think some of the limitations on RWE are, if you’re thinking about market size, how much value can you really drive? It’s clearly not zero, but it’s clearly not 10s of billions of dollars in revenue either. That was the struggle for us, how big can you be?
You deserve credit for building a significant healthcare technology company and realizing a great outcome for your shareholders.
I think we did a great job within the bounds of what was possible. Had there been more room to run from a growth standpoint, then maybe we wouldn’t have sold it. We were rational and reasonable about the fact that this gets harder, as you get bigger, to keep the growth rate up. We thought Roche was a great long-term owner because they have decade-long horizons, and how they think, they’re willing to make long-term bets, and they understand the space. That turned out to be true, they’ve been fantastic. A lot of our decision-making was around the market size.
You’re also been an active investor in healthcare. Are there any opportunities in health tech that are particularly interesting to you now?
The #1 thing I’m seeing that’s interesting in healthcare services is that there are clearly some care delivery models that are better done virtually. There are a few mental health and particular substance abuse disorders, anything to do with teenagers, and even possibly physical therapy, where it actually is a better experience for patient care to be done virtually. In those cases, you’re going to see some really interesting businesses get built. The margin structure has to make sense and it actually has to be a better solution. If you’re just trying to just shift it to virtual because you can, it’s not gonna work.
Tell us about Curie Bio?
The simple theme here is, that it’s all about founders. In particular, helping current and future founders of therapeutic companies get started. We have a thesis that there is an opportunity for significantly more founder-led, and founder-controlled therapeutic companies out there over the next decade. Technology has made the process of doing some early experiments faster and cheaper. The vendor ecosystem has matured to where you don’t have to do absolutely everything in-house, you can do some things virtually, you can do the whole thing virtually. There are some fundamental shifts going on in the cost and complexity of getting started as a new therapeutics company.
It is still very tedious for founders to tap into and access that ecosystem, and that’s a problem. I think it’s a societal problem as well. We are under-pursuing good ideas because the risk-reward of starting a new therapeutics company as an individual is kind of out of whack. The risk and operational work you have to do and the complexity are very high. The potential economic reward, or the perceived economic reward, is much lower than it should be. When those two things are inverted, which they are in software, it’s much easier to get started and the rewards are unbelievably high.
People will always surprise you with how entrepreneurial and creative they are. It’s remarkable how creative people can be if you can support them in the beginning and that’s what Curie is going to be focused on. We want to enable 10x the number of shots on goal.
Zach elaborates Curie.Bio in the video below. We wish Zach and the Curie Bio team the best of luck as they partner with the next generation of founders!