00:00all right guys we're gonna get started
00:01sorry for being late
00:03so I have up here Elizabeth
00:06irons is a dr. elizabeth iron sure you
00:09are professor Elizabeth irons so
00:12Elizabeth is a cancer biologist by
00:14training you got your PhD in cancer
00:16biology from University of London and
00:18you did your postdoc and then became an
00:20assistant professor which are still at
00:22at University of Miami School of
00:24Medicine got this correct cause so in
00:272011 she started science exchange then
00:30called the bench which was an online
00:34marketplace for outsourcing science
00:35experiments and she was part of the Y
00:37Combinator summer 2011 batch and
00:41thinking might have been like the very
00:42first biotech e-type company in YC right
00:45that's where we stood it we started with
00:46you and she's also the chairman of
00:49reformer therapeutics which just went
00:52through YC as well last batch she's also
00:54a part-time partner at YC which means in
00:56her free time she helps all our biotech
00:59companies which this point is like
01:02hundreds I don't know what the exact
01:03number is but hundreds of biotech
01:04companies so today I want to spend time
01:07talking about two things in particular
01:08one is running a marketplace company
01:10since that's what science exchange is
01:12and two is getting your perspective as
01:14someone who's now advised hundreds of
01:17biotech companies what you think about
01:18what's going on the space and I'm in
01:20particular because I think there's a
01:22there's an increase in slope of by of
01:25scientists coming in to the field so
01:28that'd be cool to talk about so to get
01:30started you're like I said you're a
01:33cancer biologist by training you were I
01:36believe doing breast cancer research
01:38right after college I guess why how did
01:42you even I mean you're in kind of going
01:44to down the academic route like why how
01:46did you even come up on startups where
01:48did that idea even come come from yeah
01:51yeah it was kind of random especially to
01:54start a technology company like science
01:56exchange rather than a biotech so I was
01:59actually so I guess taking my
02:02combinators advice without knowing about
02:04Y Combinator's advice to solve your own
02:08was looking at the way that scientific
02:11research was evolving and I was becoming
02:13much more specialized much more
02:15multidisciplinary hit deformities
02:17collaborations with other labs in order
02:19to access you know all of the latest
02:21technologies and I experienced that
02:22myself and realized that it was
02:25incredibly inefficient and so that
02:27process of how you find somebody to work
02:30with how you evaluate the quality of
02:32their ability to do that specialized
02:34work and then even just the
02:36infrastructure to actually work with
02:38them so in in biotech in and scientific
02:41research the actual ownership of the
02:43scientific results is very important and
02:44so just figuring out who is actually the
02:47owner of those results who has the
02:50intellectual property publication rights
02:52all of those things is it's very
02:54difficult and so I was talking about
02:57this with my co-founder and we looked at
03:00other examples of industries they'd
03:03solved a similar problem so we're
03:05looking at Oh deerskin Elance at the
03:06time which have now become up work but
03:09that was really an example of you could
03:11solve this with a marketplace of experts
03:13and that was the basis of science
03:15exchange and it's never really kind of
03:17evolved much more beyond that in terms
03:19of the fundamental goal of what the
03:21company wants to achieve can you dive a
03:24little bit deeper there so before if I
03:27wanted to do some research and run an
03:29experiment basically what it sounds like
03:31is this big disorganized process if I
03:33need ten different things super
03:35fragmented where this stuff is you I
03:38guess I did give an example of related
03:40to the research you were doing on how it
03:42was so disorganized and then how science
03:44exchange now collapses all that into one
03:46place yeah definitely super disorganized
03:50lots of fragmentation both in terms of
03:52who you work with but also how you find
03:55the information and in some cases the
03:57information is actually not really
03:59available so at the time when I was
04:01working as a breast cancer researcher I
04:04was running multiple different types of
04:06experiments and one example was
04:08microarray analysis which is definitely
04:11showing my time away from the lab
04:14because that's a very old-school
04:17but at the time it was you know what we
04:19used to analyse gene expression
04:20profiling and we just didn't have the
04:24types of arrays that I wanted to use and
04:26so I was trying to use Google and asking
04:29my friends and asking my old lab can we
04:31use your infrastructure and it was just
04:33incredibly inefficient also actually the
04:36the inefficiency drives a lot of
04:39inefficiency and pricing in the market
04:41so that was something that was a big a
04:43big learning point for me at that time
04:47point was you could actually create a
04:48company that had margin in that space
04:50because the market was so inefficient
04:52due to the lack of information so for
04:55example I was seeing literally ten times
04:59price differences for an equivalent
05:01experiment because people really just
05:04didn't have access to what should it
05:06actually cost and we see that still on
05:08science exchange today is that when
05:10people go and look for multiple options
05:12and they actually don't have the
05:14barriers of being able to work with them
05:16they see very frequently price
05:19differences that are quite extraordinary
05:20for the same results that could be
05:23generated got it so just to take one
05:28like when you were an academic I don't
05:31know if that's a bad word but you know
05:33when you're an academic maybe some
05:35people like was that a tough choice to
05:40just make that switch because I mean you
05:41spent so many years studying for this
05:42one thing and then you just are starting
05:46essentially what is a tech company that
05:48is not you know solving that one thing
05:50that you suspend you open till that
05:52point your entire life going after yeah
05:55and actually at the time I didn't really
05:58have a lot of role models to sort of
06:00look to either so I was in Miami which
06:03was not a very startup place and now
06:05it's actually it's interesting to see
06:06the evolution not just in the biotech
06:08world but just generally in the startup
06:11ecosystem and all of the infrastructure
06:13that's been created like I think
06:15University of Miami even has now an
06:17incubator associated with the school of
06:19mids and it helps people kind of create
06:21companies I remember at the time having
06:23this idea for a company and I
06:26actually asked the university because I
06:29was a little bit worried if I have this
06:30idea while I'm still working there will
06:32they try to get some ownership stake out
06:34from the idea because actually when you
06:36work in academia the the results you
06:39generate don't belong to you so I think
06:41that's kind of a common misperception
06:43that you know universities actually own
06:45your work so it's often a challenge to
06:47spend those results out of the
06:49university so it's kind of worried well
06:51if I have this idea will will they own
06:54it and so I went and talked to the tech
06:55transfer office and they were just like
06:57whatever I think go do whatever you want
07:00and it was it was kind of interesting
07:03and they didn't really have any guidance
07:04on how to do that but coincidentally I
07:07read about Y Combinator in Wired
07:09magazine so there was actually a Wired
07:11magazine interview all those years ago
07:13about Y Combinator and that was where I
07:16first learned about it and I was like
07:18wow they're sounds super cool why don't
07:20we just like apply for it and we were
07:22total outsiders we didn't know anybody
07:24at Y Combinator we didn't know anybody
07:26that had gone through Y Combinator and
07:28we just applied with this idea and
07:30before it should enough to get in so you
07:33talked oh you talked a little bit about
07:35tech transfer I mean I do think that a
07:37lot of people like me who are thinking
07:39about starting a start-up do you think
07:41that is a huge issue and in some sense
07:44there are from some universities we've
07:46have to work with what some of our
07:47startups it is a huge issue what's your
07:49do you have any advice for them on how
07:51to navigate that process like should
07:53that be the barrier in your mind yeah I
07:57do think it is an issue if particularly
08:00for scientific status because if you're
08:02trying to spend our actual research that
08:04was done already and I think people
08:08often ask me like what's a good time to
08:09join Y Combinator if you're creating a
08:12science startup and actually I do think
08:14a good time to join us once you've
08:16already generated a lot of the R&D from
08:19grant funded work because you don't want
08:22to spend years and all of this money
08:25trying to you know pursue a research
08:27hypothesis you really want to be in the
08:29development phase and so I think that
08:33gets very difficult with a lot of the
08:34universities I think some of them are
08:39sort of found a friendly and create ways
08:43to more readily license the results that
08:45were generated but it's still you know
08:47it's still a significant obstacle how do
08:53you see that academics is a trend
08:55they're moving more into startups or
08:57what do you see or maybe we just get a
09:00the bubble of you here in in San
09:02Francisco but I mean that's our only the
09:04thing I see anecdotally but is that do
09:06you think that's a growing trend overall
09:08yeah I think people are more willing to
09:12just create companies I think generally
09:15like think that's sort of a macro trend
09:17way beyond Sciences you know really just
09:21entrepreneurship has become this viable
09:24career path and people are like oh yeah
09:25I could be I could create a company and
09:27it's just become a lot easier like the
09:28infrastructure that's being created
09:30through you know Y Combinator but also
09:33through things like stripe and things
09:35like AWS has made it so much easier for
09:39people to create companies and I think
09:41in the science space we're starting to
09:43see that so obviously Sciences changes
09:45you know one way to do that but there's
09:48some really interesting other evolutions
09:51like lab central and other lab space qb3
09:53that provide you know lab space by the
09:56bench that's very cost effective and you
09:58can just rent you know one bench per
10:00month and be able to get started really
10:03cost-effectively so those things are
10:05lowering the barriers to entry and so I
10:07think academics what I would like to see
10:10more of is how can we allow you know the
10:13true I would say the analogy to what
10:15we've seen in the in the engineering
10:18world the software engineering world is
10:20that the actual developers the software
10:22developers become the entrepreneurs and
10:25they maintain that and really build this
10:27company around them I think that's still
10:30very difficult for scientists
10:33particularly PhD and postdoc scientists
10:37so they actually make the discoveries in
10:39these labs so the other ones doing the
10:41work a lot of the time for them to
10:43actually be the ones to be the founder
10:46of the company is pretty challenging and
10:48usually what you'll see is a
10:50well-established famous PI's
10:54principal investigator kind of being a
10:56co-founder of that company just to give
10:58it legitimacy and then that usually
11:00allows people to bring in funding but
11:03you know I think that's also kind of sad
11:04like why do why do we have to have these
11:07sort of figure heads who own large
11:09stakes of these companies in order for
11:12them to get off the ground and get
11:14started and I want to see that change
11:15you that's something we as you know talk
11:18a lot about to people who when they
11:21apply and then we see that kind of
11:22structure and then you try to go help
11:23them fix it essentially don't wanna lose
11:27too much of your company too soon
11:29all right so science exchange so you
11:32already described you know what problem
11:34you're trying to solve and then you've
11:37decided at this point in 2011 to start
11:39the startup how did you get your first
11:41users and you have two sides in the
11:42marketplace so it's doubly hard for you
11:44how did you approach both sides yeah I
11:47think marketplaces are really
11:49challenging because you don't control
11:52really the success of the business so
11:54you don't I mean you do but you don't
11:56have you're not just kind of building
11:59something and selling it you have to
12:01then make sure you have the things to
12:03sell on the other side so it is really
12:06challenging I think when we first
12:07started we we were very MVP like in our
12:12approach so I think in the first version
12:14of Sciences change which as you
12:15mentioned was actually called the bench
12:17which was renamed during Y Combinator
12:19thanks to programs very good advice
12:21about our poor naming choice so you know
12:25we we kind of threw up this website and
12:28it was let you like put your what you
12:31need on this website and then other
12:33people will see it and they will give
12:35you you know options which is even at
12:38the time as a scientist I was like I
12:39wouldn't use this make it seem like you
12:42know very quick to get something up and
12:44see who would test it out but it was
12:46obvious right from the start they would
12:48actually have to create you know a true
12:50marketplace and have supply that was
12:53visible and and actually even even more
12:58so with Sciences change what we had to
12:59do is create a curated marketplace that
13:02operates in the b2b sector so what we
13:04had to do is build QA so Quality
13:07actually qualifying contract every
13:10provider that's available through the
13:12marketplace and then surface all of the
13:15results so that when somebody comes to
13:17the marketplace they can literally find
13:19what they need and they can give them
13:21the information that's required for them
13:23to be able to quote accurately because
13:25they already know there's
13:26confidentiality agreements in place and
13:28and then they actually also had to
13:31manage those projects because these
13:32aren't just put it in your cart and
13:34check out you have to actually manage a
13:36project with somebody who works in
13:38another part of the world from you and
13:40so all this all this had to be built
13:41into the software so that was many
13:44iterations of learning about where we're
13:46really as the value that comes from what
13:49we are providing and I think
13:51marketplaces have evolved a lot from you
13:54know being predominantly consumer
13:55focused there's a lot of interesting b2b
13:57examples that are emerging in size
13:59exchange obviously you know sits very
14:02firmly in that category and well always
14:06more difficult supply or demand
14:09definitely demand so supply for us is it
14:15was pretty straightforward in the sense
14:17that you know providers are looking for
14:22work and they providers of these
14:24services they is a very fast growing
14:26industry so also highly fragmented and
14:29they actually also suffer the same
14:30challenges that the demand size has so
14:33for them if they want to sell one of
14:35their projects they actually have to go
14:38on contract with you know each client
14:40they work with which can take several
14:41months and they have to go through
14:43Quality Assurance and gets it up and so
14:45for them if they are on science exchange
14:47now all of a sudden every one of our
14:49clients can work with them instantly so
14:52that's you know that's a huge value
14:53proposition it took us a lot longer to
14:56convince the large players in the market
14:58so the large CRO is that this would be
15:01something that they should be part of
15:03but actually we have been able to do
15:05that as well so I think one of the
15:07important lessons that we learned
15:08quickly was for this to be a really
15:10effective solution it had to be the
15:13platform that was used for everything at
15:15the companies that were using it so not
15:17just one-off come and shop on the
15:19marketplace but actually use
15:21a system for managing all of the
15:23projects that are going on with the
15:24external partners and in order to do
15:26that you truly do have to have all of
15:28the major players that they want to work
15:31with on the platform and I've seen other
15:33companies who have b2b marketplaces
15:36struggle a little bit where they haven't
15:38had the large established players on the
15:41supply side and without that I think
15:43it's very difficult to to have an
15:46enterprise partnership that uses a
15:47platform do you remember the very first
15:50on the demand side the very first
15:51customer what were they yeah I remember
15:55actually so Alphaeus customers were
15:57totally my friends who were who were
16:00scientists who I would tell you if you
16:02have experiments you're running you have
16:04to try out science exchange and and they
16:07actually you know was great user
16:09feedback because for them it did turn
16:11out to be you know it's really high
16:13touch very you know very concierge
16:16service so they found great options that
16:18were really cost-effective so it wasn't
16:19like you know they just used it because
16:21we made them but I think they they
16:24taught us you know enormous amount but I
16:27also remember the first customer that if
16:29it was just you know not related to us
16:31like just came on to the platform and
16:33used it and that was pretty exciting
16:35we're actually in Italy on it my
16:38brother's waiting and I was like oh my
16:39god somebody who's that website yay stop
16:42the wedding I need to check this out
16:44only you know what was the experiment
16:47they were trying to run do you remember
16:48it was microarray as well like this was
16:50popular at that time
16:51oh so we talked a lot about that sort of
16:56school product market fit for you what
16:58was that like when did you know you had
17:01that and how did you define it in what
17:02metrics were you looking at yeah for us
17:04I think we defined product market fit
17:06when we've had our first large
17:09pharmaceutical client actually use the
17:12platform for whole sectors of their
17:15business so for us that was am jian and
17:17they used they launched science exchange
17:19for all of their discovery research
17:21which was a really big deal for us at
17:25and we sort of realized oh we actually
17:28have something that you know they do a
17:30lot of steps manually at the moment and
17:31this part this product actually
17:33automates you know automates providing
17:35all the information about who they
17:37should work with who who is you know all
17:41of these new options for new types of
17:43technologies that they might be just
17:45exploring and then even actually their
17:47business and business intelligence
17:49around you know looking at spend and
17:52supplier performance so those things
17:55those things took us a while to kind of
17:57figure out where the core value props
17:59were but you know you're onto something
18:02when a big company who spent many years
18:04and money trying to build the problem it
18:09solve that problem internally has all of
18:11a sudden just shifted to your product
18:13and it's starting to pay you for that
18:14that's a big thing not just in biotech
18:17what in general for yeah I think yeah if
18:19I mean much more so I think in software
18:21there's it's really interesting because
18:24people always ask me today like who are
18:26your main competitors and you know how
18:29do you kind of think about competitive
18:30differentiation and all that and still
18:32to me like our main competitor is status
18:35quo so that people still sit up these
18:38you know really crazy SharePoint
18:41processes to manage their external
18:44vendors and I always go in there I'm
18:46like this is horrible like why are you
18:48doing this and and you know big
18:52companies they just love SharePoint
18:53they're like oh yeah my SharePoint site
18:55I have created all of these like really
18:58Byzantine processes which only apply to
19:00like five vendors I work with and
19:02everyone else has to go through some
19:04other crazy process and I'm like okay
19:06but you can just transition all of that
19:08onto this platform and you know it's
19:10always this interesting kind of battle
19:13of who set up that process because
19:15trying to convince them that you know
19:17the software automates all that hard
19:19work is often actually one of our
19:21biggest challenges so one of the things
19:24I really love that you're doing and
19:26announced a long time ago it was the
19:27reproducibility initiative so the idea
19:30that it seems obvious to do which is you
19:32know if you run an experiment and you
19:34publish results someone else should be
19:36able to reproduce it to validate
19:38that it's being done it's it so can you
19:40talk a little bit about why you're doing
19:42that what you're doing and why has this
19:44never been done is it just been a money
19:46problem or is it something that only
19:48science exchange actually can do because
19:50you have the process in place yeah the
19:53reproducibility initiative is I still
19:56think it's a really cool project it's
19:57still very controversial even years
20:00later Lee oh yeah I said come to my god
20:02still so controversial it's
20:04controversial because of a lot of
20:06reasons I think the main reason is that
20:09people are not sure of the value of it
20:12which i think is legitimate like I think
20:14there's still a question of how
20:17efficient is it to replicate experiments
20:20and that's that was actually you know
20:22when you say is that science exchange
20:24was the only way we could do it like elf
20:26our hypothesis was that yeah having a
20:29network where people already had all of
20:31the you know write essays or the right
20:33infrastructure or the right animal model
20:35would be the only way that was actually
20:37practically possible to do it
20:40previously obviously people didn't have
20:42that they had that same challenge of oh
20:43I have to go find somebody who has that
20:46same animal model or has that same
20:48instrument and then I have to convince
20:51them and contract them and all that kind
20:53of stuff so yeah the reproducibility
20:56initiative actually is kind of science
20:59exchanges part of our mission to improve
21:02the quality and efficiency of scientific
21:04research and that that project is quite
21:07broad so it covers both like things like
21:10antibody validation which we've done a
21:12lot of Surrey agent validation things
21:15like reanalysis of Epidemiology results
21:19so we worked with the Gates Foundation
21:20to do that and then the projects that
21:23we're most well-known for which is
21:25actually replicating published results
21:28and we do that both for the
21:29pharmaceutical industry which is a way
21:31easier and way less controversial
21:33because they just come and say hey I'm
21:35interested in this result and then we
21:37quickly find you know a facility that
21:39can replicate the key results and
21:41provide this back to them however the
21:44the cancer biology reproducibility
21:46project and the prostate cancer
21:47foundation project are the two ones that
21:50people have followed the most which is
21:52applications where we've actually
21:53published the replication studies and
21:56that's controversial because people
21:58don't generally do this so I think it's
22:01it goes against the cultural norm of of
22:04really not publishing replication
22:06studies I also think people are really
22:10afraid and they sort of mix up failed
22:13replication with things like fraud so
22:17people are really afraid that if the
22:18result is is found to be non
22:20reproducible that that will have a
22:23negative impact on their career which
22:26which i think is you know probably true
22:30but it shouldn't be so the reality is
22:32that the results that have been
22:34generated show that most published
22:37results are not reproducible so if
22:39that's true then we should try to
22:42understand that and study it which is my
22:44opinion as a scientist like we should
22:46study the science behind why is this the
22:48case rather than try to go after
22:50individual people and say like oh your
22:53science is horrible because clearly
22:55that's not the case clearly it's that
22:57most things are not reproducible when
22:59they publish because of a variety of
23:00complex factors of which we've started
23:03to unpacked but I think there's still a
23:05lot of work to do what are like the one
23:07are the top two complex factors so I
23:10think the main reason like that I've
23:11seen things and not reproducible is the
23:14quality of SA validation so it's really
23:18interesting when you go and talk to
23:19pharmaceutical companies because they
23:21also transfer experiments so they
23:25replicate experiments in the sense that
23:27they'll see it something up at their
23:28facility and then they'll outsource to a
23:30CRO to run their si for them over and
23:33over again and they're they call it tech
23:35transfer or si transfer and that process
23:37is incredibly efficient they can do this
23:40very easily and it's got a high
23:42probability of working and so when you
23:44look at that I didn't think it's much
23:46different than if you take a published
23:48result and you try and rerun that si in
23:50another lab but the main difference is
23:52that when you work in a pharmaceutical
23:55company you have SOPs you
23:57of everything documented you look at sa
24:01variability you look at positive and
24:02negative controls there's just a lot
24:05more actual validation of the type the
24:08actual experimental essay that's used
24:09before it's transferred and that's very
24:12different than in academia so in
24:14academia you tend to take oh yeah I've
24:16got this animal model in my lab or I've
24:18got this you know cell line in my lab
24:19I'm just going to run this experiment
24:21and hopefully you include a positive and
24:24negative control but maybe not and you
24:26also don't look at things like
24:28variability of the sa reproducibility of
24:30the sa and I think a lot of what we see
24:34in publish results is likely to be noise
24:36rather than true experimental effects
24:39and so until we kind of understand that
24:41more we will have an issue with
24:44reproducing those results that's sort of
24:46scary you're saving the world so save us
24:50from this problem so going back to
24:54science exchange what is the biggest
24:56challenge as a CEO of a marketplace
24:58company that you have today and how is
25:02that how is that has that challenge
25:04morphed over time like what was the
25:07biggest challenge when you just started
25:08- what is the biggest challenge today oh
25:10yeah I'm sure like so many challenges
25:13along the way when we first started
25:17science exchange the biggest challenge
25:18was can we get anybody to use it you
25:24know that was kind of the first
25:25challenge in the end can we can we build
25:28something we are the very conservative
25:31pharmaceutical industry will use it and
25:33that was important for us because when
25:35you look at spend break down the market
25:39is you know predominantly for outsourced
25:42research focused in the pharmaceutical
25:43industry so so that they control over a
25:48hundred billion dollars of spins per
25:50year and so if we can if we couldn't get
25:52them to use the platform we would have a
25:54lot of issue with really creating a
25:56truly large company today I think our
26:00biggest issues are more around scaling
26:02so how do we we're still only eighty
26:05five people we've we've just signed a
26:08couple of extremely large partners
26:10and so we are really trying to execute
26:13against some pretty tight deadlines with
26:16staff that are focused on you know just
26:20one of these large integrations and so
26:22trying to kind of make sure that we
26:24don't take on too much while also
26:27realizing there's enormous opportunity
26:28so we don't want to sort of overlook
26:30that so I think just staying focused
26:32executing on the right things and being
26:35really disciplined about that which is
26:37always really hard when you have when
26:39you have multiple opportunities that you
26:41can go after I think that discipline is
26:42incredibly important yeah focus is
26:45really hard especially as you get more
26:47employees and you get more funding and
26:49stuff like that so I want to switch
26:51gears a little bit and talk about
26:52biotech in startups as we talked about
26:54earlier there has been this implosion of
26:56biotech startups recently especially in
26:59why do you why do you think that is
27:01where are the causes for this
27:03yeah the biotech industry is so
27:06interesting right now and it's actually
27:08amazing I love it's not just here in
27:10Silicon Valley it's everywhere so the UK
27:14Cambridge San Diego you know everywhere
27:18there's like lots and lots of biotech
27:21startups and and I think it's it's a
27:25really cool time for that sector of the
27:27industry it's I think driven by a couple
27:32of things one capital so access to
27:34capital is you know never as
27:36unprecedented and the amount of capital
27:39that's going into biotech is definitely
27:41unprecedented there's also there's also
27:45you know a sort of interesting evolution
27:49and in terms of where biology is in
27:51terms of the actual therapeutic
27:53modalities that are available so I'm not
27:55sure that people sort of who outside the
27:57industry may not kind of recognize this
27:59as a real turning point for biotech so
28:03up until very recently there was really
28:06only small molecule inhibitors and then
28:09you know more in then after that there
28:12was Genentech and mg and so there was
28:14actual biologic so antibodies and
28:17proteins but just in the last year we
28:19had gene therapies approved we had cell
28:23approved we had our nai approved is
28:26actual marketable products for the first
28:28time so all of a sudden you have this
28:30huge window opened up for you and the
28:33different types of approaches you can
28:36take that are viable to create a
28:38commercial product so I think the
28:40science is caught up to a point where
28:42there's just some phenomenal
28:44opportunities to tackle diseases in a
28:48way that was not possible previously
28:50interesting sort of like Moore's Law and
28:52biotech yes really it's a really really
28:54exciting time and there's also I think a
28:58convergence on you know really a
29:01critical mass of people looking at okay
29:04I have this experience in the
29:05pharmaceutical industry I'm going to go
29:07into the biotech industry so again like
29:10just so many people with really great
29:13scientific experience with true
29:15experience of bringing not from academia
29:17actually in industry bringing drugs to
29:19market and now going into biotech and
29:21they're taking that risk and founding
29:24companies and being early employees and
29:26these like five person companies and I
29:28never had seen that previously I think
29:29that's a new phenomenon as well so you
29:32just in this cool intersection of youth
29:34built the software company marketplace
29:36also a company and that intersects with
29:39biotech and then on the same at the same
29:41you have the perspective of advising
29:43lots of more true I guess biotech
29:45companies what do you what's the
29:48difference between biotech and software
29:50it's so different in terms of running
29:52with the startup itself yeah it's really
29:55different and I think in some ways you
30:00know or let me let me rephrase the
30:02question how is it the same as are there
30:04any similarities I think it's the same
30:06in terms of people and focus funding all
30:11of those things that kind of same so
30:13like trying to find the right people and
30:15retain them building a good culture all
30:18of those things actually I think YC is
30:19really being able to do a great job of
30:22having companies across different
30:24sectors because of that focus on really
30:26learning from successful entrepreneurs
30:28so big lessons about creating a company
30:31and not so much about the very you know
30:33my new tactical details bio takes a
30:37because you can't really change the
30:40outcome of the science like the science
30:42is the science so either it works or it
30:43doesn't work and that's very different
30:46than software where you can actually you
30:48know pivot around and keep developing
30:50your product and keep getting part of
30:53market fit with with science is about
30:57key milestones that demonstrate you know
31:00real inflection points in terms of
31:03mitigating the risk of the drug working
31:05so actually having like different
31:08different stages along the development
31:10path where you know you've avenged got
31:13to a point where you demonstrate in a
31:15clinical trial that your drug is
31:16effective at you know treating the
31:19disease that you're trying to treat what
31:21does an MVP mean in biotech yeah I think
31:26MVP and biotech is as I don't even I
31:32don't know if there is any MVP so you
31:35know for a biotech company that's in the
31:37therapeutic space you know really you're
31:39not going to have any commercial revenue
31:42until you have an approved product that
31:45is sold on the market which actually
31:47most biotechs never have so most bio
31:50ticks go through you know process of
31:53developing new therapies and they they
31:56do partnerships with large companies
31:59that fund you know the expensive
32:01clinical development of those products
32:03and so that's really we are a lot of
32:05them end up kind of exiting so they
32:07either acquired or they know sell or
32:10partner that product in early clinical
32:12development we're starting to see which
32:15I think is really cool some biotech
32:17companies actually commercialize their
32:19own products now and that that was
32:21something that didn't happen for quite a
32:24long time so just recently there's
32:26several companies that are actually you
32:29know selling their product themselves so
32:31building their own sales force and
32:32distribute distribution channels and
32:34those those challenges are really
32:38interesting mostly those companies are
32:41going after rare diseases where you can
32:43you know go after key opinion leaders
32:45and have a really efficient sales
32:47channel but still it's still really
32:49exciting to see them do that
32:51that's great well let me actually double
32:56down double-click on that point you made
32:58so there's essentially what you're
33:00saying is that I can bring my product to
33:03market without getting acquired which
33:05used to be the case at that specific
33:09juncture what is what has changed that
33:11is allowed that's happened yeah that's a
33:14good question I think capital so access
33:17to capital are really important that the
33:19companies can actually access enough
33:21capital to get all the way through to
33:24approval I also think the FDA has done
33:27some really interesting work trying to
33:30come up with reasonable clinical
33:32development strategies for particularly
33:34rare diseases where companies can
33:37actually get registration with a fairly
33:39small trial so that makes it actually
33:42feasible for a small company to be able
33:44to do that and then in terms of
33:47distribution I do think that strategy of
33:50just building really strong patient
33:52advocacy networks working with the key
33:54opinion leaders in the community through
33:56the hospitals that these patients are
33:58treated at provides a way to distribute
34:00that in the past I think if it's a
34:04blockbuster indication it becomes very
34:07challenging that you have to build a
34:08huge sales force you have to go and kind
34:11of sell this drug with you know sort of
34:14old-school pharmaceutical sales reps
34:16it's not it's not easy to do that I also
34:19see this proliferation of startups maybe
34:23actually since vein of signs exchange
34:25where you're helping try to speed up
34:27like experiments you're helping speed up
34:29trials you're helping speed up getting
34:32through the FDA process and so that may
34:34be helping a little bit as well yeah
34:36definitely like we've we've funded some
34:37really interesting companies in this
34:39space that kind of put in place for
34:41example regulatory infrastructure for
34:43the FDA so the expertise is out there
34:47and people are starting to productize
34:49that in a way that maybe wasn't
34:53available previously so people would
34:54have to go and hire you know regulatory
34:56consultant that would be very expensive
34:58you know literally like more than
35:00hundred thousand dollars to come up with
35:01come up with your FDA's
35:04and now there are companies that you can
35:07actually go and and work with them to
35:11provide the the productized vision of
35:13that process so enzyme is you know the
35:16company that that Y Combinator funded
35:18what you think is super interesting and
35:20all of these are just enabling
35:22technologies that will hopefully provide
35:25the infrastructure similar to what we've
35:27seen in the selfie space so you know
35:31like we said before advised and you've
35:33mentored hundreds of biotech founders at
35:35this point what are some of the common
35:37mistakes you've seen biotech founders
35:40make and what are the ones that you
35:42should you just need like what do you
35:45avoid at all costs because it's a
35:47detrimental yeah I mean I think like our
35:52bio tech founders first of all are
35:54really amazing because they are often
35:57those people who have taken that risk
36:00and have kind of stepped out of academia
36:02or other very established careers and
36:06see it I'm going to go do the startup
36:08and there isn't really that many you
36:10know role models for me to follow or
36:11like success stories for me to follow so
36:13I think they're they themselves are
36:15incredibly impressive and often I do
36:17offer sales with them and I'm like I'm
36:19just learning from them like it's
36:20actually really really cool but I think
36:22you know one mistake I have seen people
36:25make is just in it's it's I think all
36:28startup founders kind of do this is like
36:30not doing the killer experiment like not
36:32actually just cutting to the chase and
36:36these are just the minimal things I need
36:39to do to truly like answer the question
36:41and you almost don't want to do it
36:43because if it doesn't work the company's
36:45kind of did and I sort of a lot of the
36:48time push our companies a little bit I'm
36:50like but why haven't you done that
36:51experiment like that experiment would
36:52tell you you know today if this is going
36:56to work or not and I think it's just
36:57it's hard when it's your own company to
36:59kind of do that but the sooner you do it
37:02the more the sooner you have enough
37:03money to kind of work on something else
37:06because like we talked about with
37:07science if you can't get the actual
37:09science to work you you know really are
37:11in a difficult position so say you're a
37:15scientist or you're not
37:17who's looking to get into startups I
37:20think a lot of people still think maybe
37:24you do you need some kind of business
37:25co-founder let's help me do XY and Z do
37:28you what do you think about that in you
37:30know should that be a goal no I don't
37:34think that it should be a goal I think I
37:37actually do have a business co-founder I
37:39think it would probably be mad that I
37:41say that but and he's great and I
37:43literally couldn't have felt sizes
37:45change without him but but not because
37:48of his business background right like
37:49because he's a great co-founder because
37:52he's like hustles and figures stuff out
37:54and we work well together I think
37:58business so much of it is just common
38:00sense like you I used to get so you know
38:04consumed that I didn't have this finance
38:05background I didn't know how to read
38:07like all of the income statements and
38:09all that and then I realized after a
38:11while okay I'm just gonna sit down with
38:13my business co-founder and he's gonna
38:15teach me and so he taught me and then I
38:16was like oh it's so obvious like it's
38:18not it's definitely not rocket science
38:21so I think I think that the business
38:24side you know you really as a CEO and
38:28when you grow the company one of the key
38:30skills is actually recognizing the areas
38:32that you're good at and better and what
38:34where you should be investing and
38:36bringing in top talent and so for us we
38:39we recently actually hired a CFO so a
38:43year ago we hired a CFO and that's been
38:45I think good for the company but also
38:48good from a perception perspective so as
38:50the company reaches a growth a growth
38:53stage actually having that sort of
38:55legitimate CFO person does help you but
38:59from like a starting out I mean starting
39:01out you should just find people who
39:03really want to solve the same problem as
39:05you and really care about it and also
39:07who you really like working with because
39:10that's by far the most important thing
39:12on the flip side if you're not if you're
39:16not a scientist and let's say you're a
39:19programmer but you're interested in
39:20getting into biotech what what should
39:24you do like should I go back to school
39:26get my PhD like what's the what's a
39:29potential path for me
39:30there's a good question you know part of
39:34me thinks like you should go back to
39:36school like I think there is this
39:38there's a lot of interest from Silicon
39:42Valley in biotech like people are super
39:44interested in just like even just
39:47hacking themselves like this whole kind
39:48of movement around like you know really
39:51personalized like understanding all of
39:53your own biology I think it's really
39:55cool and I by the way do think that the
39:58future of biotech and we are I always
40:01think about for reformer and and the
40:04work that we are doing is I believe that
40:06the future will definitely involve a
40:08strong component of user pays so I think
40:11the products that are developed will
40:13have to be in indications that the
40:16patients are actually willing to pay for
40:18and there's a lot of research at the
40:21moment that's in areas which you know I
40:23think are potentially problematic
40:26because they're sort of diseases where
40:28people are not really that sick or they
40:30don't really feel sick and so getting
40:32them to adhere to those medications very
40:35very challenging in contrast like things
40:37like migraine drugs actually amgen's
40:39migraine drug has outperformed its
40:41predictions in the market by ten times
40:43and I think that's because people
40:45genuinely go to the doctor because they
40:47debilitated by migraines and they will
40:50pay for those drugs so so I think trying
40:53to you know think about ways that you
40:55can focus on on users is good so anyway
40:58to go back to three strings it got so
41:00Jay about should computer programmers
41:03what should they do I think for biology
41:05and actual scientific research there is
41:07this element of just getting in the lab
41:09and truly understanding how experiments
41:11are designed and how to interpret them
41:12which I don't I don't know that you can
41:14just learn from not doing but then there
41:19is like a lot of use for for example
41:21bioinformatics and other analysis tools
41:24and platforms where people can get
41:26involved without having lab experience
41:27and we do have some successful companies
41:30and Y Combinator that are founded by non
41:33scientific founders that are in the
41:35biotech space so notable labs is one
41:37that I think is incredibly impressive
41:40the founders have basically self-taught
41:42themself everything about
41:43you know the sector that they're in and
41:45the super smart and hungry and you know
41:48straight away when I interviewed them
41:49was like yeah they know just as much as
41:52you know PhDs who work in the space so
41:55two more questions one is just going
41:58back to science exchange for a minute
41:59looking back of all decisions you've
42:03what what's in the early days what was
42:06aside from just starting the start of
42:08itself that's obviously very critical
42:09but what's like a decision you made
42:12where you're looking back you're like
42:13that that was a game-changer that was an
42:15inflection point in my business Wow
42:18there is a good Christian game-changer
42:23actually I think the decision to do the
42:26reproducibility initiative was a
42:28game-changer and it was non obvious so
42:30in some ways the reproducibility
42:33initiative goes against focus it was
42:36kind of a distraction like okay we're
42:38gonna do this project but it's not
42:40directly related to just growing the
42:41marketplace although the marketplace was
42:43used to run the project but it was so
42:47timely and so high-profile that it did
42:50change the branding and the
42:52opportunities for science is change in a
42:53way that we never expected and it did
42:56open all of the doors that eventually
42:58led to our pharmaceutical partnerships
43:00to really like a lot of the 60s of
43:03science exchange so that was probably
43:04one example interesting did not know it
43:07it kind of when I saw it I was like oh
43:09wow science exchange I knew it could be
43:11big but it could be like this much
43:13bigger because it just showed the you
43:15know what you could do with it okay last
43:17question is always my favorite question
43:21in a hundred years from now I mean
43:23you're only you've been around for eight
43:24years at seven eight years now but in a
43:26hundred years from now what do you think
43:28science exchange will be yeah there is a
43:31such an interesting question because I
43:33think if you just think about what the
43:36world will be like in 100 years from now
43:38I'm not sure any of us have a good
43:40answer but you know I do think that a
43:43hundred years ago the scientific method
43:46existed and people were you know doing
43:48scientific research and I said and I
43:51think scientific research will exist 100
43:55so science is change has always been
43:57extremely purpose driven so you know the
44:00company's purpose is to enable
44:01scientific breakthroughs through
44:03connections and so I think whatever the
44:06world looks like at that time that's you
44:09know what science to change will be
44:10doing and and hopefully we're just
44:13providing that infrastructure that
44:14enables people to instantly work with
44:17whoever they need to collaborate with in
44:19order to make these scientific
44:20breakthroughs happen it's crazy to think
44:23the scientific method was only
44:25discovered invented or do you want to
44:26call it just not long ago and which
44:30creates this explosion of science in
44:32general okay cool well that's all the
44:35questions I have any questions from the
44:36audience back there so the question is
45:01about whether we should look at new
45:04approaches to instead of just looking at
45:07correlations to look for causality
45:13especially in a space with unknown
45:15unknowns so you know I think the way
45:18that so I'm a biologist by training so I
45:21tend to think that you know we're not
45:24just looking at correlations we try to
45:26design experiments that allow us to
45:30change something in the system there is
45:32a controlled system and then read an
45:35output from that and determine whether
45:37that fits of our hypothesis that we
45:40changed something and it has therefore
45:41had this downstream impact I think
45:45there's you know really interesting work
45:47that's been done on correlations
45:50particularly with real-world data so
45:53trying to look for ways that we can
45:55actually use humans in the wild to
45:58examine new theories that we have and
46:02then apply those back into the lab but
46:04basically of all when you're in the lab
46:06you are kind of using model systems to
46:09try to reduce the unknown unknowns so
46:13that you can test specific theories so
46:33the question is about going from an idea
46:35to an MVP in a short space of time and
46:38the steps that required to do that and
46:40for us actually there was one of the
46:42lessons that I really took when I
46:44started science exchange was try to do
46:47something really quickly and get it off
46:50the ground because I see a lot of people
46:52start companies and they have a lot of
46:54enthusiasm at the start and then they're
46:56also working like full-time jobs and
46:58trying to do this on the side and the
47:01progress you make obviously is limited
47:03because you just don't have the time to
47:06put into it so we are possible I think
47:08it's it's really great if you can kind
47:10of just take time out to say okay I'm
47:12going to do like Y Combinator or
47:14something for three months and really
47:16so for us we literally had the idea and
47:19then it was the I think it was February
47:222011 and and we kind of we're talking
47:26about different ideas and then we we
47:28thought this is really a good idea and
47:30so then we applied to Y Combinator so we
47:34made this video and it was just me and
47:36my co-founder we had nothing and then
47:38actually Y Combinator Alexis Ohanian he
47:41Skyped me and he said you're not gonna
47:44get in and I was like oh no why it was
47:46slow because you don't have a technical
47:48co-founder and you really need a
47:49technical co-founder and so we then
47:52spoke to all of our friends and we found
47:54a technical co-founder and this was all
47:57just in like two weeks and then we built
47:59actually a really hack to give the
48:01vision of the MVP and we came to our
48:04interview and we had already got
48:06something kind of was very basic but we
48:10had something and and then we got in and
48:14we moved out here in May and it was
48:17three months of just okay now let's get
48:20that launched and we were actually doing
48:21transactions off platform so we actually
48:23were talking with all of our scientists
48:26and I was traveling a lot talking to you
48:29know all of the people who I could get
48:30to use the product so we actually did
48:32hundreds of thousands of dollars of
48:33transactions during that time just to
48:36prove that we understood the demand and
48:38the supply side you know yeah so the
49:23question is about the importance of
49:25credentials in the biotech space and I
49:28think credentials are very important so
49:30if you can have you know credentials in
49:33the space it's just you know it's going
49:35to give you obviously a huge advantage
49:37but in saying that I don't think it's we
49:40have examples where people have been
49:42successful without that and how they
49:45were successful is by being incredibly
49:49credible themselves when you actually
49:51interview them so in the example of
49:54notable labs when I interviewed Matt and
49:56Pete they just had researched everything
49:59about the space like they'd read every
50:01scientific paper they knew in depth
50:03about cancer stem cells about the
50:05limitations about the essays that they
50:06wanted to use and I met with them much
50:09longer than I would have would have if I
50:11had a surrogate which would be the
50:13credential of them having a PhD from a
50:15top university but by talking to them it
50:18was clear that they did understand the
50:20space and that they were incredibly
50:22motivated due to a family connection to
50:26put something in place that could help
50:28solve this issue which was in the case
50:30of them they were looking for new
50:32therapeutics for glioblastoma and and so
50:36I think if you are not a scientist
50:39having a personal driver of why you're
50:42doing this actually can serve as a
50:44surrogate to get you in the door so it
50:47can get you meetings with top scientists
50:49it can get you meetings with patient
50:51advocacy groups that can help you sort
50:54of get the company started I think it
51:01can I think if you're building a biotech
51:03company you obviously have to build a
51:05scientific team and then you'll end up
51:08with PhDs in your team but you can be a
51:11co-founder without a PhD so the question
51:27is about when did I decide to leave the
51:29university and so I was so fortunate
51:33when I started science exchange and I
51:34think a lot of people don't realize how
51:38difficult and I totally don't take this
51:40for granted like when people ask me
51:41about our journey of studying sizes
51:43change I think we had you know enormous
51:46luck in the sense that my boss was the
51:49Dean of Medicine at the University of
51:52Miami and he was incredibly supportive
51:54of science exchange so he thought it was
51:56a great idea he thought that if you know
51:59if I didn't do it then like somebody
52:02else would do it and so he actually let
52:04me take three months off to go and do
52:07this and he looked after my lab for me
52:08while I was gone and then once we're out
52:11here it was clear that the idea was
52:13going to be successful and we raised
52:15funding straight out of YC and so I
52:18decided I'm not going to go back and I
52:20was actually nervous about telling him I
52:23wasn't going to go back but he was so
52:25amazing about it he was just like yeah
52:27and that was doing great like I knew it
52:28would be a great success and so having
52:30that mentor who gave me the opportunity
52:33I just I think not many people get that
52:36especially in academia like that's
52:38actually again where
52:39sometimes have this frustration of I
52:41hear the opposite of PhD students and
52:44postdocs they tell me oh my my boss just
52:47really didn't want me to leave like
52:49really didn't want me to start a company
52:51actively worked against me
52:53rather than helped me and I think about
52:55my experience and how different it would
52:56have been if I didn't have his support
52:59all right last question Rena yes so the
53:16question is about quality control and a
53:17two sided marketplace so for us quality
53:20control is incredibly important and
53:22actually is one of the core value
53:24propositions of science exchange so we
53:26qualify all suppliers before they're
53:29available through the marketplace and
53:30then we also have a continuous
53:33monitoring process where we actually
53:35look at performance of every single
53:37transaction so we have more data on
53:40performance than anybody else and we can
53:43actually say with certainty well at
53:46least most indeed than other people this
53:48provider will likely do a very good job
53:50on this type of experiment we also put
53:53in place the way our actual platform is
53:57structured it is clear and there clear
54:00outline of the deliverables that are
54:02generated the expectations are set up
54:04front and I think a really interesting
54:06step that we track closely as science
54:08exchanges Net Promoter Score is 78 and
54:11our suppliers net promoter score is 67
54:15and the industry average is zero so we
54:19think that's amazing because it's the
54:21same suppliers but when used through the
54:24platform they perform much better and I
54:26think the reason is because it's
54:28structured and it's clearly outlined
54:30what's going to be delivered and then
54:31there's any expectation that if you
54:33don't perform the information will
54:35actually be available to everyone else
54:36when they're making a decision so it
54:38becomes a very strong and Center for
54:41people to perform and make sure that
54:42they're delivering what they agreed upon
54:45all right thank you so much Elizabeth