A Conversation with Elizabeth Iorns - Advice for Biotech Founders

All right, guys, we're gonna get started. Sorry for being late. So I have up here Elizabeth Irons. Is it Dr. Elizabeth Irons? No, you're Professor Elizabeth Irons. So Elizabeth is a cancer biologist by training. You got your PhD in cancer biology from the University of London, and you did your postdoc and then became an assistant professor, which you’re still at at the University of Miami School of Medicine. Got this correct?

So in 2011, she started Science Exchange, then called The Bench, which was an online marketplace for outsourcing science experiments. She was part of the Y Combinator Summer 2011 batch. I think it might have been like the very first biotech e-type company in YC, right? That’s where we stood it. We started with you. She's also the chairwoman of Reformer Therapeutics, which just went through YC as well last batch. She's also a part-time partner at YC, which means in her free time she helps all our biotech companies, which at this point is like hundreds. I don’t know what the exact number is, but hundreds of biotech companies.

So today, I want to spend time talking about two things in particular: one is running a marketplace company since that's what Science Exchange is, and two is getting your perspective as someone who's now advised hundreds of biotech companies on what you think about what's going on in the space. I’m in particular because I think there's an increase in the slope of scientists coming into the field, so that'd be cool to talk about.

So to get started, you're, like I said, a cancer biologist by training. You were, I believe, doing breast cancer research right after college, right? How did you even, I mean, you’re kind of going down the academic route. Why and how did you even come up on startups? Where did that idea even come from?

Yeah, it was kind of random, especially to start a technology company like Science Exchange rather than a biotech. So I was actually, I guess, taking my combinator's advice without knowing about Y Combinator’s advice to solve your own problem. So I was looking at the way that scientific research was evolving and I was becoming much more specialized, much more multidisciplinary, had to form collaborations with other labs in order to access, you know, all of the latest technologies. I experienced that myself and realized that it was incredibly inefficient.

So that process of how you find somebody to work with, how you evaluate the quality of their ability to do that specialized work, and then even just the infrastructure to actually work with them—so in biotech and scientific research, the actual ownership of the scientific results is very important. Just figuring out who is actually the owner of those results, who has the intellectual property, publication rights—all of those things are very difficult.

I was talking about this with my co-founder, and we looked at other examples of industries that had solved a similar problem. So we were looking at Odesk and Elance at the time, which have now become Upwork. But that was really an example of you could solve this with a marketplace of experts, and that was the basis of Science Exchange. It’s never really kind of evolved much more beyond that in terms of the fundamental goal of what the company wants to achieve.

Can you dive a little bit deeper there? So before, if I wanted to do some research and run an experiment, basically what it sounds like is this big disorganized process. If I need ten different things, it’s super fragmented. Could you give an example related to the research you were doing on how it was so disorganized and then how Science Exchange now collapses all that into one place?

Yeah, definitely. Super disorganized. Lots of fragmentation, both in terms of who you work with, but also how you find the information. In some cases, the information is actually not really available. So at the time when I was working as a breast cancer researcher, I was running multiple different types of experiments.

One example was microarray analysis, which is definitely showing my time away from the lab because that's a very old-school technique now. But at the time, it was what we used to analyze gene expression profiling. We just didn't have the types of arrays that I wanted to use. I was trying to use Google and asking my friends and asking my old lab, “Can we use your infrastructure?” It was just incredibly inefficient.

Also, the inefficiency drives a lot of inefficiency and pricing in the market. So that was something that was a big learning point for me at that time—point was you could actually create a company that had margin in that space because the market was so inefficient due to the lack of information. So for example, I was seeing literally ten times price differences for an equivalent experiment because people really just didn't have access to what should it actually cost. And we see that still on Science Exchange today, is that when people go and look for multiple options and they actually don't have the barriers of being able to work with them, they see very frequently price differences that are quite extraordinary for the same results that could be generated.

Got it. So just to take one step back, when you were an academic, I don’t know if that’s a bad word, but you know when you're an academic, maybe some people, like, was that a tough choice to just make that switch? Because, I mean, you spent so many years studying for this one thing and then you just are starting essentially what is a tech company that is not, you know, solving that one thing that you spent your whole life going after up until that point.

Yeah, and actually at the time, I didn't really have a lot of role models to sort of look to either. So I was in Miami, which was not a very startup place. And now it’s actually, it's interesting to see the evolution, not just in the biotech world, but just generally in the startup ecosystem and all of the infrastructure that's been created. Like, I think the University of Miami even has now an incubator associated with the School of Medicine, and it helps people kind of create companies.

I remember at the time having this idea for a company and I actually asked the university because I was a little bit worried, “If I have this idea while I'm still working there, will they try to get some ownership stake out from the idea?” Because actually when you work in academia, the results you generate don’t belong to you. So I think that's kind of a common misperception that, you know, universities actually own your work, so it's often a challenge to spin those results out of the university. So I was kind of worried: “Well, if I have this idea, will they own it?”

I went and talked to the tech transfer office, and they were just like, “Whatever, I think go do whatever you want.” It was kind of interesting, and they didn't really have any guidance on how to do that, but coincidentally, I read about Y Combinator in Wired magazine. So there was actually a Wired magazine interview all those years ago about Y Combinator, and that was where I first learned about it. I was like, "Wow, that sounds super cool. Why don’t we just apply for it?" We were total outsiders.

We didn't know anybody at Y Combinator. We didn't know anybody that had gone through Y Combinator, and we just applied with this idea. Before it showed enough to get in.

So, you talked a little bit about tech transfer. I mean, I do think that a lot of people like me who are thinking about starting a start-up do you think that is a huge issue? And in some sense, there are from some universities we’ve had to work with; what’s your do you have any advice for them on how to navigate that process? Like should that be the barrier in your mind?

Yeah, I do think it is an issue, particularly for scientific startups because if you're trying to spend our actual research that was done already. I think people often ask me, "What's a good time to join Y Combinator if you’re creating a science startup?" And actually, I do think a good time to join us is once you've already generated a lot of the R&D from grant-funded work. Because you don't want to spend years and all of this money trying to pursue a research hypothesis. You really want to be in the development phase, and so I think that gets very difficult with a lot of the universities.

I think some of them are trying to be more sort of founder-friendly and create ways to more readily license the results that were generated, but it's still, you know, it’s still a significant obstacle.

How do you see the trend of academics moving more into startups, or what do you see or maybe we just get the bubble of you here in San Francisco? But I mean that’s the only thing I see anecdotally. But is that do you think that's a growing trend overall?

Yeah, I think people are more willing to just create companies. I think generally, that's sort of a macro trend way beyond science. Really just entrepreneurship has become this viable career path. People are like, "Oh, yeah, I could create a company." It's just become a lot easier, like the infrastructure that's being created through, you know, Y Combinator, but also through things like Stripe and things like AWS has made it so much easier for people to create companies.

I think in the science space, we're starting to see that. So, obviously, Sciences changes one way to do that. But there's some really interesting other evolutions like Lab Central and other lab spaces, QB3 that provide, you know, lab space by the bench that’s very cost effective, and you can just rent, you know, one bench per month and be able to get started really cost-effectively. So those things are lowering the barriers to entry.

What I would like to see more of is how can we allow, you know, the true—I would say the analogy to what we've seen in the engineering world, the software engineering world is that the actual developers, the software developers become the entrepreneurs, and they maintain that and really build this company around them.

I think that's still very difficult for scientists, particularly PhD and postdoc scientists, so they actually make the discoveries in these labs. So the other ones doing the work a lot of the time. For them to actually be the ones to be the founder of the company is pretty challenging. Usually, what you'll see is a well-established famous PI, principal investigator, kind of being a co-founder of that company just to give it legitimacy. That usually allows people to bring in funding, but you know, I think that's also kind of sad. Like, why do we have to have these sort of figureheads who own large stakes of these companies in order for them to get off the ground and get started? I want to see that change.

That’s something we, as you know, talk a lot about to people who when they apply, and then we see that kind of structure and then you try to go help them fix it. Essentially, you don’t wanna lose too much of your company too soon.

All right, so Science Exchange. So you already described, you know, what problem you're trying to solve, and then you’ve decided at this point in 2011 to start the startup. How did you get your first users? And you have two sides in the marketplace, so it’s doubly hard for you. How did you approach both sides?

Yeah, I think marketplaces are really challenging because you don't control really the success of the business. So you don't—I mean you do, but you don't. You're not just kind of building something and selling it; you have to then make sure you have the things to sell on the other side. So it is really challenging.

I think when we first started, we were very MVP-like in our approach. So I think in the first version of Science Exchange, which as you mentioned was actually called The Bench—this was renamed during Y Combinator thanks to Paul Graham's very good advice about our poor naming choice.

So, you know, we kind of threw up this website. It was let you like put your what you need on this website and then other people would see it, and they would give you options. Which, even at the time as a scientist, I was like I wouldn't use this. It made it seem like, you know, very quick to get something up and see who would test it out.

But it was obvious right from the start; they would actually have to create, you know, a true marketplace and have supply that was visible. And actually, even more so with Science Exchange, what we had to do is create a curated marketplace that operates in the B2B sector.

So what we had to do is build QA—so Quality Assurance. We had actually to qualify and contract every provider that’s available through the marketplace and then surface all of the results so that when somebody comes to the marketplace, they can literally find what they need and can give them the information that’s required for them to be able to quote accurately. Because they already know there are confidentiality agreements in place and then they actually also had to manage those projects because these aren't just put it in your cart and check out. You have to actually manage a project with somebody who works in another part of the world from you.

So all this had to be built into the software. So that was many iterations of learning about where we're really as the value that comes from what we are providing. I think marketplaces have evolved a lot from, you know, being predominantly consumer-focused.

There's a lot of interesting B2B examples that are emerging in Science Exchange. Obviously, you know, it sits very firmly in that category. And it's always more difficult, supply or demand? Definitely demand.

So supply for us was pretty straightforward in the sense that, you know, providers are looking for work, and they, providers of these services, is a very fast-growing industry, so also highly fragmented. They actually also suffer the same challenges that the demand side has. So for them, if they want to sell one of their projects, they actually have to go on contract with, you know, each client they work with, which can take several months, and they have to go through quality assurance and get it up.

So for them, if they are on Science Exchange, all of a sudden, every one of our clients can work with them instantly. So that’s a huge value proposition. It took us a lot longer to convince the large players in the market, so the large CROs, that this would be something that they should be part of, but actually, we have been able to do that as well.

I think one of the important lessons that we learned quickly was for this to be a really effective solution, it had to be the platform that was used for everything at the companies that were using it. So not just one-off come and shop on the marketplace but actually use a system for managing all of the projects that are going on with the external partners.

In order to do that, you truly do have to have all of the major players that they want to work with on the platform, and I've seen other companies who have B2B marketplaces struggle a little bit where they haven't had the large established players on the supply side. Without that, I think it's very difficult to have an enterprise partnership that uses a platform.

Do you remember the very first customer on the demand side? What were they?

Yeah, I remember actually. So our first customers were totally my friends who were scientists who I would tell, “If you have experiments you're running, you have to try out Science Exchange.” And they actually, you know, it was great user feedback because for them, it did turn out to be designed as a really high-touch, very, you know, very concierge service. They found great options that were really cost-effective, so it wasn’t like, you know, they just used it because we made them, but I think they taught us an enormous amount.

But I also remember the first customer that if it was just, you know, not related to us, like just came onto the platform and used it—and that was pretty exciting. We were actually in Italy at my brother's wedding, and I was like, "Oh my god, somebody’s on that website!" Yay, stop the wedding! I need to check this out.

You know, what was the experiment they were trying to run? Do you remember?

It was microarray as well. This was popular at that time.

So we talked a lot about that sort of school product-market fit for you. What was that like? When did you know you had that, and how did you define it, and what metrics were you looking at?

Yeah, for us, I think we defined product-market fit when we’d had our first large pharmaceutical client actually use the platform for whole sectors of their business. So for us, that was Amgen, and they used Science Exchange for all of their discovery research, which was a really big deal for us at the time.

We sort of realized, “Oh, we actually have something that, you know, they do a lot of steps manually at the moment, and this product actually automates, you know, automates providing all the information about who they should work with—who is, you know, all of these new options for the new types of technologies that they might be just exploring.

And then even actually their business intelligence around, you know, looking at spend and supplier performance. Those things took us a while to kind of figure out where the core value props were. But you know you’re onto something when a big company who spent many years and money trying to build the problem to solve that problem internally has all of a sudden just shifted to your product and is starting to pay you for that.

That’s a big thing, not just in biotech, but in general.

Yeah, I think, yeah, much more so I think in software. It's really interesting because people always ask me today like who are your main competitors and, you know, how do you kind of think about competitive differentiation and all that. Still to me, like our main competitor is the status quo.

People still set up these, you know, really crazy SharePoint processes to manage their external vendors. I always go in there and I’m like, “This is horrible. Why are you doing this?” And you know, big companies, they just love SharePoint. They’re like, “Oh yeah, my SharePoint site! I’ve created all of these like really Byzantine processes which only apply to like five vendors I work with, and everyone else has to go through some other crazy process.”

I’m like, “Okay, but you can just transition all of that onto this platform.” It’s always this interesting kind of battle of who set up that process because trying to convince them that, you know, the software automates all that hard work is often actually one of our biggest challenges.

One of the things I really love that you’re doing and announced a long time ago was the reproducibility initiative. So the idea that it seems obvious to do, which is, you know, if you run an experiment and you publish results, someone else should be able to reproduce it to validate that it's being done.

So can you talk a little bit about why you're doing that, what you're doing, and why has this never been done? Is it just been a money problem, or is it something that only Science Exchange actually can do because you have the process in place?

Yeah, the reproducibility initiative is—I still think it's a really cool project. It's still very controversial even years later.

Lee, oh yeah, I said come to my god, still so controversial!

It’s controversial because of a lot of reasons. I think the main reason is that people are not sure of the value of it, which I think is legitimate. Like I think there’s still a question of how efficient is it to replicate experiments. And that's—that was actually, you know, when you say is Science Exchange was the only way we could do it? Like our hypothesis was, “Yeah, having a network where people already had all of the, you know, right assays or the right infrastructure or the right animal model would be the only way that was actually practically possible to do it.”

Previously, obviously, people didn’t have that. They had that same challenge of, “Oh, I have to go find somebody who has that same animal model or has that same instrument.” Then I have to convince them and contract them and all that kind of stuff.

So yeah, the reproducibility initiative actually is kind of Science Exchange’s part of our mission to improve the quality and efficiency of scientific research, and that project is quite broad. It covers both, like things like antibody validation, which we’ve done a lot of. Surrey agent validation, things like Reanalysis of Epidemiology results. We worked with the Gates Foundation to do that.

And then the projects that we’re most well-known for, which is actually replicating published results. We do that both for the pharmaceutical industry, which is a way easier and way less controversial because they just come and say, “Hey, I'm interested in this result.” And then we quickly find, you know, a facility that can replicate the key results and provide this back to them.

However, the Cancer Biology Reproducibility Project and the Prostate Cancer Foundation project are the two ones that people have followed the most, which are public applications where we’ve actually published the replication studies. And that’s controversial because people don’t generally do this.

I think it goes against the cultural norm of really not publishing replication studies. I also think people are really afraid, and they sort of mix up failed replication with things like fraud. So people are really afraid that if the result is found to be non-reproducible, that that will have a negative impact on their career, which I think is, you know, probably true, but it shouldn’t be.

The reality is that the results that have been generated show that most published results are not reproducible. So if that’s true, then we should try to understand that and study it, which is my opinion as a scientist: we should study the science behind why this is the case rather than try to go after individual people and say, “Oh, your science is horrible,” because clearly that's not the case. Clearly, it's that most things are not reproducible when they publish due to a variety of complex factors of which we've started to unpack. But I think there's still a lot of work to do.

What are like the top two complex factors?

So I think the main reason that I've seen things are not reproducible is the quality of SA validation. So it's really interesting when you go and talk to pharmaceutical companies because they also transfer experiments, so they replicate experiments in the sense that they’ll set something up at their facility and then they’ll outsource to a CRO to run their assays for them over and over again. They call it tech transfer or assay transfer, and that process is incredibly efficient. They can do this very easily and it’s got a high probability of working.

When you look at that, I think it’s much different than if you take a published result and you try and rerun that assay in another lab. But the main difference is that when you work in a pharmaceutical company you have SOPs, you have everything documented, you look at SA variability, you look at positive and negative controls—there's just a lot more actual validation of the actual experimental assay that's used before it's transferred. And that's very different than in academia.

So in academia, you tend to take, “Oh yeah, I've got this animal model in my lab” or “I've got this, you know, cell line in my lab. I’m just going to run this experiment.” And hopefully you include a positive and negative control, but maybe not, and you also don’t look at things like variability or the reproducibility of the assay. I think a lot of what we see in published results is likely to be noise rather than true experimental effects.

So until we kind of understand that more, we will have an issue with reproducing those results.

That’s sort of scary. You're saving the world, so save us from this problem.

So going back to Science Exchange, what is the biggest challenge as a CEO of a marketplace company that you have today? And how has that challenge morphed over time? Like what was the biggest challenge when you just started, and what is the biggest challenge today?

Oh yeah, I'm sure like so many challenges along the way. When we first started Science Exchange, the biggest challenge was can we get anybody to use it? You know, that was kind of the first challenge in the end. Can we build something we are the very conservative pharmaceutical industry will use? And that was important for us because when you look at spend breakdown, the market is predominantly for outsourced research focused on the pharmaceutical industry.

They control over a hundred billion dollars of spins per year, and so if we couldn’t get them to use the platform, we would have a lot of issues with really creating a truly large company. Today, I think our biggest issues are more around scaling. So how do we—we're still only 85 people. We’ve just signed a couple of extremely large partners, and so we’re really trying to execute against some pretty tight deadlines with staff that are focused on, you know, just one of these large integrations.

So trying to kind of make sure that we don't take on too much while also realizing there’s enormous opportunity, so we don’t want to overlook that. I think just staying focused, executing on the right things, and being really disciplined about that, which is always really hard when you have multiple opportunities that you can go after.

I think that discipline is incredibly important.

Yeah, focus is really hard, especially as you get more employees and you get more funding and stuff like that. So I want to switch gears a little bit and talk about biotech and startups. As we talked about earlier, there has been this implosion of biotech startups recently, especially in Silicon Valley. Why do you think that is? What are the causes for this?

Yeah, the biotech industry is so interesting right now, and it’s actually amazing. I love it. It's not just here in Silicon Valley; it’s everywhere. So the UK, Cambridge, San Diego, you know, everywhere, there are lots and lots of biotech startups. I think it's a really cool time for that sector of the industry.

I think it’s driven by a couple of things: one, capital. So access to capital is, you know, never as unprecedented and the amount of capital that’s going into biotech is definitely unprecedented. There’s also an interesting evolution in terms of where biology is in terms of the actual therapeutic modalities that are available.

So I’m not sure that people outside the industry may not kind of recognize this as a real turning point for biotech. Up until very recently, there were really only small molecule inhibitors. Then, you know, after that there were Genentech and mg, and so there were actual biologics—antibodies and proteins. But just in the last year, we had gene therapies approved. We had cell-based therapy approved.

We had RNAi approved as actual marketable products for the first time. So all of a sudden, you have this huge window opened up for you in the different types of approaches you can take that are viable to create a commercial product. So I think the science is caught up to a point where there's just some phenomenal opportunities to tackle diseases in a way that was not possible previously.

Interesting. Sort of like Moore's Law in biotech.

Yes, really, it's a really, really exciting time. There’s also, I think, a convergence on, you know, really a critical mass of people looking at, “Okay, I have this experience in the pharmaceutical industry. I’m going to go into the biotech industry.”

So again, there are just so many people with really great scientific experience, with true experience of bringing drugs to market, and now going into biotech. They’re taking that risk and founding companies and being early employees in these like five-person companies. I never had seen that previously—I think that’s a new phenomenon as well.

So you just, in this cool intersection of youth, built the software company marketplace, also a company, and that intersects with biotech. And then on the same, at the same, you have the perspective of advising lots of more true, I guess, biotech companies. What do you, what’s the difference between biotech and software?

It’s so different in terms of running with the startup itself.

Yeah, it's really different. I think in some ways, you know—or let me rephrase the question—how is it the same? Are there any similarities?

I think it's the same in terms of people and focus, funding, all of those things that kind of same. So like trying to find the right people and retain them, building a good culture—all of those things actually I think YC has really been able to do a great job of having companies across different sectors because of that focus on really learning from successful entrepreneurs.

So big lessons about creating a company, not so much about the very, you know, minute tactical details. Bio takes a different—because you can’t really change the outcome of the science; like the science is the science. So either it works or it doesn't work. That’s very different than software, where you can actually, you know, pivot around and keep developing your product and keep getting product-market fit.

With science, it’s about key milestones that demonstrate, you know, real inflection points in terms of mitigating the risk of the drug working. So actually having different stages along the development path where, you know, you've advanced to the point where you demonstrate in a clinical trial that your drug is effective at, you know, treating the disease that you’re trying to treat.

What does an MVP mean in biotech?

Yeah, I think MVP in biotech—I'm not even sure there is any MVP. So, you know, for a biotech company that's in the therapeutic space, you know, really you're not going to have any commercial revenue until you have an approved product that is sold on the market, which actually most biotechs never have. So most biotechs go through, you know, the process of developing new therapies and they do partnerships with large companies that fund, you know, the expensive clinical development of those products.

That’s really where a lot of them end up kind of exiting. They either get acquired or they, you know, sell or partner that product in early clinical development. We’re starting to see, which I think is really cool, some biotech companies actually commercialize their own products now. That was something that didn’t happen for quite a long time.

So just recently, there are several companies that are actually, you know, selling their product themselves, building their own sales force and distribution channels. Those challenges are really interesting; mostly those companies are going after rare diseases where you can, you know, go after key opinion leaders and have a really efficient sales channel. But still, it’s still really exciting to see them do that.

That’s great. Well, let me actually double down, double-click on that point you made. So there’s essentially what you’re saying is that I can bring my product to market without getting acquired, which used to be the case at that specific juncture. What has changed that has allowed that to happen?

Yeah, that's a good question. I think capital—so access to capital are really important that the companies can actually access enough capital to get all the way through to approval. I also think the FDA has done some really interesting work trying to come up with reasonable clinical development strategies for particularly rare diseases where companies can actually get registration with a fairly small trial. So that makes it actually feasible for a small company to be able to do that.

And then in terms of distribution, I do think that strategy of just building really strong patient advocacy networks, working with the key opinion leaders in the community through the hospitals that these patients are treated at provides a way to distribute that. In the past, I think if it's a blockbuster indication, it becomes very challenging; you have to build a huge sales force. You have to go and kind of sell this drug with, you know, sort of old-school pharmaceutical sales reps. It’s not easy to do that.

I also see this proliferation of startups, maybe actually since the vein of Science Exchange, where you're helping to try to speed up experiments, you're helping to speed up trials, you're helping to speed up getting through the FDA process. That may be helping a little bit as well.

Yeah, definitely! We've funded some really interesting companies in this space that kind of put in place, for example, regulatory infrastructure for the FDA. So the expertise is out there, and people are starting to productize that in a way that maybe wasn’t available previously.

So, people would have to go and hire, you know, regulatory consultants, and that would be very expensive—literally like more than a hundred thousand dollars—to come up with your FDA’s. And now there are companies that you can actually go and work with them to provide the productized version of that process.

So Enzyme is, you know, the company that Y Combinator funded.

What do you think is super interesting and all of these are just enabling technologies that will hopefully provide the infrastructure similar to what we’ve seen in the software space. So, you know, like I said before, advised and you’ve mentored hundreds of biotech founders at this point. What are some of the common mistakes you’ve seen biotech founders make? And what are the ones that you should just need? Like what do you avoid at all costs because it’s detrimental?

Yeah, I mean, I think, like, our biotech founders, first of all, are really amazing because they are often those people who have taken that risk and have kind of stepped out of academia or other very established careers and see, “I’m going to go do the startup.”

There isn’t really that many, you know, role models for me to follow or success stories for me to follow. So I think they themselves are incredibly impressive, and often I do offer sales with them. I’m like I’m just learning from them; it’s actually really, really cool.

But I think, you know, one mistake I have seen people make is just in—it’s, I think all startup founders kind of do this— is like not doing the killer experiment. Like not actually just cutting to the chase and saying, “Okay these are just the minimal things I need to do to truly like answer the question.”

You almost don’t want to do it because if it doesn’t work, the company's kind of dead. I sort of a lot of the time push our companies a little bit. I’m like, “But why haven’t you done that experiment? That experiment would tell you, you know, today if this is going to work or not.”

And I think it’s just hard when it’s your own company to kind of do that. But the sooner you do it, the sooner you have enough money to kind of work on something else because like we talked about with Science, if you can't get the actual science to work, you really are in a difficult position.

So say you’re a scientist or you’re not—who’s looking to get into startups? I think a lot of people still think maybe you do need some kind of business co-founder. Let’s help me do X, Y, and Z. Do you, what do you think about that? Should that be a goal?

No, I don't think that it should be a goal. I think, I actually do have a business co-founder. I think it would probably be mad that I say that; and he’s great, and I literally couldn’t have felt sizes change without him, but not because of his business background, right? Like because he's a great co-founder because he's like hustles and figures stuff out and we work well together.

I think business—so much of it is just common sense. Like I used to get so, you know, consumed that I didn't have this finance background. I didn’t know how to read like all of the income statements and all that. And then I realized after a while, “Okay, I’m just going to sit down with my business co-founder and he's going to teach me.”

And so he taught me, and then I was like, “Oh, it’s so obvious.” Like it’s not—it’s definitely not rocket science. So I think that the business side, you know, you really as a CEO and when you grow the company, one of the key skills is actually recognizing the areas that you're good at and better and what where you should be investing and bringing in top talent.

So for us, we recently actually hired a CFO. So a year ago we hired a CFO, and that’s been, I think, good for the company but also good from a perception perspective. So as the company reaches a growth stage, actually having that sort of legitimate CFO person does help you. But from like a starting out, I mean starting out, you should just find people who really want to solve the same problem as you and really care about it and also who you really like working with because that’s by far the most important thing.

On the flip side, if you're not a scientist, and let's say you're a programmer, but you're interested in getting into biotech, what should you do? Like, should I go back to school, get my PhD? Like what's the potential path for me?

That’s a good question. You know, part of me thinks like you should go back to school. I think there is this—there's a lot of interest from Silicon Valley in biotech. People are super interested in just like even just hacking themselves, like this whole kind of movement around like, you know, really personalized like understanding all of your own biology. I think it’s really cool, and by the way, I do think that the future of biotech and we are—I always think about for Reformer and the work that we are doing is I believe that the future will definitely involve a strong component of user pays.

So I think the products that are developed will have to be in indications that the patients are actually willing to pay for. And there’s a lot of research at the moment that's in areas which, you know, I think are potentially problematic because they’re sort of diseases where people are not really that sick or they don’t really feel sick.

So getting them to adhere to those medications is very, very challenging. In contrast, like things like migraine drugs—actually Amgen’s migraine drug has outperformed its predictions in the market by ten times, and I think that's because people genuinely go to the doctor because they’re debilitated by migraines, and they will pay for those drugs.

So, trying to, you know, think about ways that you can focus on users is good. So anyway, to go back to the three strings—it got so jay about should computer programmers what should they do?

I think for biology and actual scientific research, there is this element of just getting in the lab and truly understanding how experiments are designed and how to interpret them, which I don't know that you can just learn from not doing. But then there is like a lot of use for, for example, bioinformatics and other analysis tools and platforms where people can get involved without having lab experience.

We do have some successful companies in Y Combinator that are founded by non-scientific founders that are in the biotech space. So Notable Labs is one that I think is incredibly impressive. The founders have basically self-taught themselves everything about, you know, the sector that they’re in, and they’re super smart and hungry.

You know, straight away when I interviewed them, I was like, “Yeah, they know just as much as PhDs who work in the space.”

So two more questions. One is just going back to Science Exchange for a minute—looking back at all the decisions you’ve made, what’s in the early days? Aside from just starting the start of itself—that's obviously very critical—but what's like a decision you made where you’re looking back, you’re like that was a game-changer, that was an inflection point in my business?

Wow, that is a good question. Game-changer actually I think the decision to do the Reproducibility Initiative was a game-changer, and it was non-obvious. So in some ways, the Reproducibility Initiative goes against focus. It was kind of a distraction—like, “Okay, we’re gonna do this project,” but it’s not directly related to just growing the marketplace, although the marketplace was used to run the project.

It was so timely and so high-profile that it did change the branding and the opportunities for Science Exchange in a way that we never expected. It did open all of the doors that eventually led to our pharmaceutical partnerships to really, like, a lot of the successes of Science Exchange. So that was probably one example.

Interesting. I did not know it. It kind of when I saw it, I was like, “Oh wow, Science Exchange. I knew it could be big, but it could be like this much bigger,” because it just showed the, you know, what you could do with it.

Okay, last question is always my favorite question: In a hundred years from now, I mean you've only been around for eight years—seven, eight years now—but in a hundred years from now, what do you think Science Exchange will be?

Yeah, there is such an interesting question because I think if you just think about what the world will be like in 100 years from now, I’m not sure any of us have a good answer. But, you know, I do think that a hundred years ago, the scientific method existed and people were, you know, doing scientific research, and I think scientific research will exist 100 years from now.

So Science Exchange has always been extremely purpose-driven. So, you know, the company's purpose is to enable scientific breakthroughs through connections. I think whatever the world looks like at that time, that's, you know, what Science Exchange will be doing, and hopefully, we're just providing that infrastructure that enables people to instantly work with whoever they need to collaborate with in order to make these scientific breakthroughs happen.

It’s crazy to think the scientific method was only discovered, invented, or whatever you want to call it, just not long ago, which creates this explosion of science in general.

Okay, cool. Well, that’s all the questions I have. Any questions from the audience back there?

So the question is about whether we should look at new approaches to instead of just looking at correlations to look for causality, especially in a space with unknown unknowns. So, you know, I think the way that—so I’m a biologist by training, so I tend to think that, you know, we're not just looking at correlations. We try to design experiments that allow us to change something in the system; there is a controlled system and then read an output from that and determine whether that fits our hypothesis that we changed something and it has therefore had this downstream impact.

I think there’s, you know, really interesting work that’s been done on correlations, particularly with real-world data. So trying to look for ways that we can actually use humans in the wild to examine new theories that we have, and then apply those back into the lab. But it’s basically, of all, when you're in the lab, you are kind of using model systems to try to reduce the unknown unknowns so that you can test specific theories.

So the question is about going from an idea to an MVP in a short space of time and the steps that are required to do that. For us, actually, there was one of the lessons that I really took when I started Science Exchange was try to do something really quickly and get it off the ground because I see a lot of people start companies, and they have a lot of enthusiasm at the start, and then they're also working like full-time jobs and trying to do this on the side.

And the progress you make, obviously, is limited because you just don’t have the time to put into it. So where possible, I think it’s really great if you can kind of just take time out to say, “Okay, I’m going to do like Y Combinator or something for three months and really launch the company.”

So for us, we literally had the idea, and then it was, I think it was February 2011. We were kind of talking about different ideas, and then we thought, “This is really a good idea.” So then we applied to Y Combinator. We made this video, and it was just me and my co-founder. We had nothing.

Then actually, Y Combinator—Alexis Ohanian—he Skyped me, and he said, “You’re not gonna get in.” I was like, “Oh no, why?” It was slow because, “You don’t have a technical co-founder, and you really need a technical co-founder.”

We then spoke to all of our friends, and we found a technical co-founder, and this was all just in like two weeks. Then we built actually a really hacky to give the vision of the MVP. We came to our interview, and we had already got something kind of, it was very basic, but we had something, and then we got in.

We moved out here in May, and it was three months of just, “Okay, now let’s get that launched.” We were actually doing transactions off-platform, so we actually were talking with all of our scientists, and I was traveling a lot, talking to all of the people who I could get to use the product.

So we actually did hundreds of thousands of dollars of transactions during that time just to prove that we understood the demand and the supply side.

Yeah, so the question is about the importance of credentials in the biotech space, and I think credentials are very important. So if you can have credentials in the space, it’s just, you know, it’s going to give you obviously a huge advantage. But in saying that, I don't think it’s—we have examples where people have been successful without that, and how they were successful is by being incredibly credible themselves when you actually interview them.

So in the example of Notable Labs, when I interviewed Matt and Pete, they just had researched everything about the space; like they’d read every scientific paper they knew in-depth about cancer stem cells, about the limitations, about the assays that they wanted to use. I met with them much longer than I would have if I had a surrogate, which would be the credential of them having a PhD from a top university, but by talking to them, it was clear that they did understand the space and that they were incredibly motivated due to a family connection to really try to put something in place that could help solve this issue, which was, in the case of them, they were looking for new therapeutics for glioblastoma.

And so I think if you are not a scientist, having a personal driver of why you're doing this actually can serve as a surrogate to get you in the door. So it can get you meetings with top scientists, it can get you meetings with patient advocacy groups. That can help you sort of get the company started.

I think it can. I think if you're building a biotech company, you obviously have to build a scientific team, and then you'll end up with PhDs in your team, but you can be a co-founder without a PhD.

So the question is about when did I decide to leave the university. And so I was so fortunate when I started Science Exchange, and I think a lot of people don’t realize how difficult—and I totally don’t take this for granted—like when people ask me about our journey of studying Science Exchange, I think we had, you know, enormous luck in the sense that my boss was the Dean of Medicine at the University of Miami, and he was incredibly supportive of Science Exchange.

So he thought it was a great idea. He thought that if I didn’t do it, then like somebody else would do it. So he actually let me take three months off to go and do this. He looked after my lab for me while I was gone. Then once we were out here, it was clear that the idea was going to be successful, and we raised funding straight out of YC. So I decided I’m not going to go back.

I was actually nervous about telling him I wasn’t going to go back, but he was so amazing about it. He was just like, “Yeah, and that was doing great. I knew it would be a great success.” So having that mentor who gave me the opportunity, I think not many people get that.

Especially in academia, like that’s actually, again, where sometimes have this frustration of, I hear the opposite of PhD students and postdocs. They tell me, “Oh, my boss just really didn't want me to leave. He really didn’t want me to start a company. He actively worked against me rather than helped me,” and I think about my experience and how different it would have been if I didn't have his support.

All right, last question, Rena.

Yes, so the question is about quality control in a two-sided marketplace. So for us, quality control is incredibly important and actually is one of the core value propositions of Science Exchange. So we qualify all suppliers before they’re available through the marketplace, and then we also have a continuous monitoring process where we actually look at the performance of every single transaction.

So we have more data on performance than anybody else, and we can actually say with certainty—well at least most indeed than other people—that this provider will likely do a very good job on this type of experiment. We also put in place the way our actual platform is structured; it is clear and there is a clear outline of the deliverables that are generated. The expectations are set up front, and I think a really interesting step that we track closely as Science Exchange’s Net Promoter Score is 78, and our suppliers’ net promoter score is 67, and the industry average is zero.

So we think that's amazing because it's the same suppliers, but when used through the platform, they perform much better. I think the reason is because it’s structured and it’s clearly outlined what’s going to be delivered, and then there’s any expectation that if you don’t perform, the information will actually be available to everyone else when they’re making a decision. So it becomes a very strong incentive for people to perform and make sure that they’re delivering what they agreed upon.

All right, thank you so much, Elizabeth. You like everyone. [Applause]