- Alright, so this will be last, this'll be the fourth session we have on the earliest stage of starting up, this organizational skill one. And so today, we have Ann Miura-Ko. We're very lucky to have her. Ann is a co-founder of
a firm called Floodgate, which is one of the most interesting, I think one of the most
interesting funds around, and they invest in a
wide variety of stuff. We'll talk about physical goods, we'll talk about marketplace,
we'll talk about biology, we'll talk about math. Ann is probably the most
educated person in the room, at least, by a long stretch,
and she'll tell us about it. I'm just gonna let her jump in, and she's gonna talk about what at Floodgate they see that's interesting, and how they think about what
they see early stage that they think translates into
important, durable companies. So we'll just let Ann talk,
and then we'll do some Q and A like we did
with Michael on Tuesday. - So, yeah, my talk is called
Hunting Thunderlizards, and Thundarlizards is actually a term that my cofounding partner
has talked a lot about. His name is Mike Maples. We've been working together
for a number of years, since 2008. So just to dive in and
give you a little bit background about me. Palo Alto native, so I've grew up here. Went to Paly High, went to,
snuck into some Stanford parties as a high school senior. Went to Yale, Electrical Engineering. Went to McKinsey, and
then got into investing kind of by chance. Partner at Charles River
Ventures by the name of Ted Dintersmith gave me a shot, and we actually had this
incredible conversation about books and music that we loved, and at the end of it, he
basically offered me a job. I ended up starting working
for him a day before 9/11. So my second day of work
was 9/11, and I got to see what happens in venture investing when everything comes
to a screeching halt. I eventually ended up at Stanford Ph.D. I got my Ph.D. in the
Operations Research Group doing math modeling of computer security. Had the opportunity to interact a lot with Stanford students, and
really fell in love with the act of teaching. In fact, I'm gonna give
a plug for my class. In the Spring quarter,
I'm gonna be teaching MSNE275, and this is sort of a test run for some of the elements of the course that I'm gonna be teaching. And then I ended up at Floodgate. Floodgate, we started in 2008. I was thinking of starting
my own company at the time and I was told by one of my mentors I should go and find an angel investor who has a lot of experience looking at very early stage things and see what's going on in the real world. I started doing that
with Mike once a week. We'd look at companies together. In the midst of that, he
turned to me and said, "I think I've accidentally raised a fund. "And I think in that
case, you should join me. "This isn't a venture back startup "like you're thinking of doing, "but it's a backed venture
startup, so let's go." And so in 2008, I got started. These are some of the companies
I've been involved with. I had a really great
string of beginner's luck, so in 2008, as you can imagine, was a really hard year economically. But turns out, it was
a great year to become a venture investor,
because not many people were actually writing checks,
and so between 2009 and 10, I actually invested in three companies. My first was TaskRabbit. My second one was
Modcloth, and the third one was Zimride, which turned into Lyft. And all three of those companies, I really kinda wax nostalgic about them, because it was a period where I really had no background in investing. I had nothing to turn to to
say this is my track record. And so each one of those founders, whether it was Leah
Busque or John Zimmerman and Logan Green or Eric and Susan Koger. They, as much as I took a bet on them, they took a bet on me. And so since then, I've
invested in Refinery29, which has really taken off. Ayasdi actually was born
in this very building in the Math Department, and
so I'll talk a little bit more about why I think companies
like that are significant. And so what are thunderlizards? This is sort of a central
theme for Floodgate. We talk a lot about these animals. They are obviously
inspired, not by dinosaurs, but they are inspired by Godzilla. And they represent what we think of when we think of entrepreneurs, the tremendous scaled
entrepreneurs that have gone through blitzscaling. They are born from what we
call radioactive atomic eggs. And so there is something,
there's a genetic mutation from the day zero with
these entrepreneurs. They then swim across the Pacific Ocean and have this incredible
long journey through lots of dangerous things. They stop off in Hawaii
often, and then end up in, depending on whether Mike tells the story, he says that they end up in Tokyo. I like the version where it
ends up in San Francisco. And he emerges in San Francisco Bay with an attitude, and
starts to wreak havoc. And we like the older
versions of this as well. They shoot laser beams
out of their mouths. The best ones eat trains. Some of them will eat ships. But what does that mean? Why does that imagery evoke
entrepreneurship for me, for Mike, for people we talk
about thunderlizards to. Well, we think about it in
sort of four different areas. One is, and probably this
is the most powerful notion, is that they've somehow
converted their advantage to truly disruptive power. And when I say disruptive
power, that to me evokes a lot of what we see in Godzilla. You don't see this in
Godzilla, but these companies will minimize technical
and organizational risk, or organizational debt. These are the things
that happen in a company as you're acting quickly, where
you are accruing some cost. And great companies
will minimize that debt. You will achieve product market fit. So all of these companies will have an incredible feel for the product, but will also have a pull for
their product from the market. It's almost uncontrolled
pull of their product. And then lastly, this is
something that Peter Thiel has also talked about,
they avoid competition. And so these are the four
things that we believe result from the fact that
you are a thunderlizard. The thunderlizards are really rare. And this is actually
a slide from McKinsey. And I thought this was really fascinating because if you took all
of the IPOs that happened between 1980 and 2012,
there's about 3,000 of them. But the number that
was most striking to me was how many of these have more than $4,000,000,000 in revenues. Well, it turns out it's only 17 companies. And if you can imagine what
it takes to get to be one of those 17, those are the
truly legendary businesses. And even getting to an IPO
is already exceedingly rare. There are roughly only about
25 companies in any given year that exit at greater than $500,000,000. And so if you think about
that, that is the actual when you have your exit event, there's only about 25 of
those in any given year, and it really doesn't
matter what year it is. You might have a slight
blip one year, but this year looks like it's gonna be
right around 20 to 25 again. And so when we think about how do you describe what that looks like? So we know that they're exceedingly rare, but Floodgate, we invest
either really early or way too early, so we
are the first institutional financing that happens
for a startup company. We're cutting checks between
$500,000 and $3,000,000 into a company. And so that means that we need to spot it way earlier than most other
larger venture capital firms. And so Mike and I spent a
lot of time thinking about not only what do we try to spot, but also as we work with these companies, what's the groundwork
that needs to be laid so that when they go on to
their next round of financing or future rounds of
financing, they don't accrue the kind of debt that could really hamper the company's growth. And so we think about it in terms of what I call the Value Stack. And this is really sausage being made. I describe this as sort of
stuff that I'm still working on. This is the first time actually teaching this in a course format. I've whiteboarded it previously, so if any of you have questions on this, I'd love to get feedback as we go along. If I start at the bottom, I
call it proprietary power. I used to call it technology. And so this is the place
where you get something very unique, it's an insight that you have that then allows you to avoid competition. You go to the next layer,
and it's product power. That's where you achieve this
magical product market fit. And then you get to a point where you have that product market fit, but now you need to make sure that you have a real sustainable company. And then ultimately from that company, you don't wanna just have a
great sustainable company, you want to have one that dominates, and that's what we call category kings. And so if you can translate
the initial energy that you have within
your company as a startup into all of these
different levels of power then we believe you have the capacity to become one of these thunderlizards, and we're looking for early signs of that. I'm gonna go into each of
these little levels of power. So within the bottom of
the value stack, yeah. - [Student] When you
look at a new company, is there something that rates
them on those four metrics, or are those the foundation by which you, lenses by which you look at a company? - Yeah, so I'll repeat the question. Do I actually have sort of a scorecard, essentially is the
question, along the lines of all of these different dimensions. We don't actually have a scorecard, but these are different levels
that we will talk through. And we will pretty explicitly talk about what the company has, and
what the company doesn't have. And oftentimes when
turning down a company, it's because there's something in there that we feel there's lacking
on one of these dimensions. So if we start with proprietary power, this is probably the part
that I spend the most time, because this is probably the most evident at the early stages of a company. And this is, it could
be a technical insight. So oftentimes, it's gonna
turn up as proprietary IP. So Ayasdi's story is that it started out with Gunnar Carlsson, who was the Chairman of the Math Department here at Stanford. He had 25 years of research in topological data analysis, and his graduate student actually turned it into real product. And from there, they
published four papers. I encoungered Gurjeet at that time, and then he turned it into a business plan and ultimately, we
funded it at that moment. But it was 25 years of research, so when you look at
Ayasdi, one of the things that's interesting about
it is they have technology that is in many ways unassailable. Another way to look at it is do you have access to scarce supply. I provide DeBeers as a canonical example of you have access to diamonds, then you're probably
in a pretty good spot. And so scarce supply can
actually be a great advantage. The creation of high switching costs, that's Keurig, right. So if I buy a Keurig
machine, then I'm probably gonna keep buying those
little Keurig cups, because I already bought
the Keurig machine and I want my coffee. So that's a really good
locking in mechnanism. Network effects, well
you're being taught by Reid Hoffman, who is probably
the king of network effects, so I won't say much about that, but obviously that also
creates incredible power and a lock-in for both
the consumer, but then from the network, then
you drive even more power, and then more people come. The authentic team, when I thought about authentic team, the immediate story that came to mind for me was Lyft. And the reason why was
when Zimride first came to pitch for me, they
had this incredible story of how the two founders met. And so I was looking through the notes from the last lecture,
where Michael Dearing was talking about transportation, and I love that part of his lecture, because it was actually very reminiscent of the pitch that John and Logan gave me. When John talked about
why he joined Zimride, so Logan was first to get it started, and he had this interesting story of how he had joined the Transportation Board of Santa Barbara, because he was convinced there was gonna be something there, and then we turned to John and said, "Well how did you get interested?" He told us an incredible story of how when he was at Cornell,
he had this professor who did essentially Michael Dearing's talk about transportation, and
there was this one slide that he put up where he said, "Look, whenever there's been a huge "discontinuous shift in the economy, "it's always been lead by something "that's changed in transportation." He talked about the canals,
he talked about trains, then he talked about the highway system. And John said, "I know that
there's something next, "and I want to be part of that," And that's what Zimride's about. And what's interesting to me is Zimride then switched into Lyft. That story, I've shown that pitch, the pitch for Zimride to other people, and you wouldn't actually know until you got to the product description that the pitch was not for Lyft. And so we really look at the
authenticity of the team. So when I invested in Zimride in 2010, it took two years for them
to get to the product pivot, which got them to Lyft. And so that was two years
of quite a bit of work, quite a bit of heartache
to get to that point, and you can imagine what
kind of impact that has, and so we have a strong viewpoint that all of this, proprietary power, can be unlocked pretty quickly in
the early stages of startup, and something that's probably
easiest for us to assess. Any questions on that? Yep. - [Student] In today's
world, where you think about how the barriers to entry
have lowered significantly, 'cause technically people
being able to work on projects, how has that made you
change the way you think or place certain emphasis on different aspects of this project? - So all the building product is easier. Actually, developing
these insights is not. And so I think that's
what's important to me is there's something that you
know that no one else knows. One of the things that was told to me as I was doing research
for developing this stack was the founder of Viva
Systems, whose company IPOed a couple years ago, they had
only raised $7,000,000 total, and I think four of it was touched. But he said, "Every great
startup has one fundamental "assumption that has
a less than 50% chance "of being correct, "but if true, will give you a "20x, 100x advantage in market." And I love that kind of notion of what is non-consensus, but right for these magical startups? For Viva Systems, it was actually buidling their entire platform on top of force.com. Lot of people thought
that was just a crazy, kind of stupid idea. Why would you build, how
would you possibly build a billion dollar company
on top of sales force? He proved that he could. So regardless of how easy
it is to build a company or build a product, we think that having some sort of insight
that says, "Well I have "access to very scarce supply," or "I have proprietary IP,"
or "I know something about "the market shift that's happening "that no one else does." That's what enables you
to unlock the market power before anyone else does. Any other questions? Alright. Then we go to product power. And what's interesting is with product, it's being able to achieve
that product market fit. But before you get here, so if you have proprietary IP, there's
a good chance that you never actually get to product power, and we see examples of that
probably all over Stanford, which is you find this really
interesting piece of research and you're trying to figure out how to turn it into a startup company, and then someone at
some point says to you, "That's technology in
search of a problem." And so you might have
incredible proprietary power, but it can be very difficult
to turn into product power. Product power is when you
achieve product market fit. And it's, and a lot of people
throw out product market fit as if, it's a product that
hits any kind of market, someone in the world likes it, then you have product market fit. That's not product market fit. Product market fit implies market power. It's the market that needs
to be large and growing. Oftentimes when you
hit product market fit, it feels uncontrolled. In fact, with Instagram,
there's a great example of that. Mike talks about, at the very first week when they launched Instagram,
they had this moment where they realized that, and they knew, they had the sense that on the weekends, the traffic
would suddenly increase, and he knew, it was on
a Monday or Tuesday, that if they didn't get
the Instagram service off of servers and into the cloud, that they were completely
screwed for the weekend. And so they worked day and
night to get that to happen, and they were finally able to get it to be a scaleable service by Friday. And he knew that at that
point, he said within 12 hours of having launched
Instagram, they actually had a really good sense that Kevin Systrom had a very good sense that
this was gonna be a huge hit and I think Mike was a little
bit more skeptical for a while but he always felt like it was a train whose wheels were coming off. And so the question I
often get from founders is do I have product market fit? And the answer to that question is if you're asking, you don't. And then, a lot of people will also ask isn't it true that you can
actually create a market if you have a transformative product? And that's true, too,
but it rarely happens. I think the example that I
could come up with was VMware. VMware was so transformative
that it created an entire market, where there
really wasn't one previously. Any questions, there? Yeah. - [Student] Are there
maybe smaller metrics that maybe you would see
clues in the beginning? 'Cause for example, Airbnb. I know that they didn't really get that skyrocket thing at the very start. They worked for probably a couple years before they really caught
that product market fit. So are there smaller clues
that you could see beforehand? - I feel like that was
actually, even with Airbnb, there was this moment
where, I think it was the Democratic National
Convention, where Obama first presented. That was a real moment for them. There is actually a moment where you start to see that grow. I think the place where it
can be harder is marketplaces. And marketplaces can actually take a really, really, really long time, because it's this very difficult balance of the supply and the demand. And oftentimes, you'll see companies kind of getting that balance wrong in the early days, and
it takes a little while for them to really mesh. But I think of this as
really a binary concept. And there can be early signs. It really depends on what you are, what kind of company you are, what those metrics are. Any other questions? Yeah. - [Student] Do you know of any good ways to test the market before
going into it full force? - Yeah, so customer discovery. I've taught a lot with Steve Wang. There's the whole sort of process
and four steps to epiphany and what Eric Ries talks about. I think there's good
ways of just going out to customers and testing
out your hypotheses. Oftetimes, actually,
I've seen entrepreneurs track all of their hypotheses, even in an Excel spreadsheet, and start to say, "This is my hypothesis. "This is how I'm gonna test it, "and this is what I expect to see." And it's very much a scientific method. I think the thing where that can go wrong is that it's as much an
art as it is a science. You have to know when your
experiments were wrong so that the experiment
itself could be flawed. It could also be that the
way in which it was presented wasn't quite right. And so you have to be
fairly honest with yourself about what went wrong in those tests, but I do think that
there are ways in which you can actually access the customer, get in front of them and see
if there's an aha moment. Again, with Instagram, it was interesting. They said when they were with Burbn, they noticed actually that
you could post pictures, but it was like a five step process. You had to go into
settings, you had to find this bespoke email address, and then you would have to email your pictures to this bespoke email address,
and then it would be somehow posted into Burbn, but people were actually doing that. And they noticed within that stream that that was something
they were interested in and that's why they doubled down on it. And so oftentimes, it's
not what you want to see and what you're looking
for, but the things that are sort of outside of realm that help you notice, "Oh that's a problem "that I might wanna solve." And so oftentimes when we say, when you go into do customer discovery, you don't want to say,
"Take a look at my product "and tell me what you think." You want to ask them about, "What's your life like? "What are some of the
problems that you face?" And it's in those conversations that you might actually find inspiration for the real product
that you want to build. Anything else? Okay. And then from there,
we go to company power. This is actually where, when
you're scaling your company, you're talking about discovering
a scaleable business model. You're trying to figure
out what the cultural values are for your company, and what your identity is as a company. It's how you develop your
own company narrative, the personality of the place. It's how you develop
talent, reward talent, provide career paths for individuals, and it's how you foster communication within that culture. And companies that I think have done this really, really well, Facebook. But also Netflix is really well known for some of their HR documentation. But it goes way more than, far beyond what I would say is just HR, and I would emphasize, do you have that scaleable business model,
does it make sense today? I would argue that there's
lots of unicorns today that are still searching for
that scaleable business model. And that means that there's a lot of debt that can actually be incurred. If you don't have the
right processes in place, you will actually incur
a lot of technical debt. You will also incur a lot of
company organizational debt. I see this in my own companies, where you have eight C-level executives and you're thinking to yourself, "How do we get here?" Well, it's just organizational debt. And then the compensation
schemes are really messed up. That's organizational debt. People feel like they
don't have a career path. That's organizational debt. And then so these are also, in some ways, when you look at the founder
and you talk to them, you can also start to spot
are these people who are thinking about these
things, who will be open to thinking about these
things, who want to think about these things,
and how open are they to that kind of conversation? And those are things that
we actually do look for in the early stages as
well, because again, if we believe that the founder is gonna be the CEO for the long
term, these are things that they're gonna have to grow into. And we want to at least
be aware of whether or not they care about these things. Yeah. - [Student] Can you talk
more about the signs in the early stage that
you look for in founders, 'cause often this when you're scaling, and you're looking at embryos. So is there anything you sort of see as patterns from people who
care about this later on? - Yeah, I think one of the
things is actually early hiring. So how good are they at hiring, and what are they willing to give up to get the best people? We have one company, actually, that has been able to attract
really ridiculous talent from big companies,
Facebook, Square, Twitter, and you talk to them about
how he was able to to that. He spends a lot of time thinking about not only who he's hiring and
what kind of networks they'll bring to the table, but also
how they do interviewing. What do they put together
for a compensation plan? They think about it in a
really rigorous manner, almost just as much as they
think about the product. And to me, that's, those are early signs. When I invested in
Ayasdi, part of the thing that convinced me was that
Gurjeet brought on board this incredible product person, and he was very open about the fact that there were things that he did not know, and that he wanted to
know, but he also wanted to surround himself with people
who would support him in that. And so all of these types
of signals are around. What is the type of
company you want to build? What's your fundamental values? A lot of the founders that are successful are fairly generous with equity. They want people to feel like an owner and not just an employee. So those are the types of
things that we're looking for. It's more philosophy
rather than explicit things that you'll actually see as signs. It's not metrics driven. And then lastly, category power. From my friends at Play Bigger, which is a great organization, two guys who are CMOs of companies. But they talk about how do you derive category kings. And to me, what is it? These are the guys that convert all of these advantages that they have and then create this
incredible disruptive power. And in a lot of senses, they will actually define a whole new category, because they don't want to compete. They want to be the big
Godzilla on the block. And so I think a great example
of that is Netflix, actually. When they came in, they weren't saying, "I'm gonna compete with Blockbuster." They didn't compete, they
created a whole new category and then destroyed Blockbuster. They destroyed the category and rebuilt it in a way that suited them. And I think Starbucks
is another great example that who would've thunk that people would want to drink $3 coffee, a $5 coffee when you could get it for $.50. They completely redefined
the buying behavior and enabled a new category to emerge. Amazon's done the same thing,
Apple's done the same thing. And when we look at companies, one of the things that we
think is really powerful is how does the founder think about the language that defines the
area that they're going into and how do they position their
company within that market? If they feel like they're
not in control of that market and they're allowing the
market to define who they are, that's something that
worries us a little bit more. We would prefer to see companies be more proactive about it. And so those are the four elements of the value stack that we look at. And obviously, this is
actually probably the hardest to view at the stage that we invest in, but in a lot of ways, it can
actually be the most important. If you can become a category king, that's how you capture value the most. Yep. - [Student] Can the category power ever come before the company power, or is that a bad sign if that happens? - No, I think it can. You can start defining a lot of that. It's separate in some ways. Category power in a lot
of ways is external, whereas company power is internal. It's how you look
internally at your business. It's how you build that
sustainable business model. Some people will try to
get to category power by buying their way in,
so there are companies that will spend a ton of
money on customer acquisition and just buy a ton of customers, and try to claim category power, when in reality, it makes no sense from a unit economic standpoint. And so my argument is, at that stage, those are both very important. You need to actually make sure you have a good, sustainable business model. And there must be another
way that you can actually get to category power. - So the question we like to ask everyone, and part of this is because
Reid is so good at it, is what can you safely ignore? So the pitch, one way to
interpret the pitch is, well you should be good
at a lot of things, including building a big company and a big category with no technical debt and no organizational debt. And so at the stage of sort of zero, or 1% to 15, what can you ignore? And then I'll ask how much
can you really ignore it? - Yeah, I think it's
more what do you focus on first and foremost? How do you prioritize? For me, the priority is at
the bottom of that stack. What is that unique advantage
that you're building today, that proprietary power? So if it's IP that you
have, how strong is that? How do you build that
into an increasing mote? If you have a supply chain advantage, how do you really make
that into a product? I think the second piece you can't ignore is the product, product market fit. How do you get there? Those are probably the two things that I focus on, alongside the team. How do you develop the team internally? Zero to 10, you don't
really need to think about career path at that point. You don't need to think
about titles so much. So the organizational debt probably comes a little bit later. You could actually start
to incur technical debt, but a lot of times, you don't even know, it's more important to be aware of what the technical debt is, rather than seeking to
really fix everything before you move on. 'Cause you don't, at
the stage that we're at, you half the time don't know what's gonna stay and what's gonna go, right? And so I would say the
one thing that I would say is the most important for us is the speed at which you're making these decisions. When a founder is frozen
or just needs more data, that's the first sign
of distress that we see. - Yeah, and that was a
thing that came up with Sam and Michael both, is speed of
decision making is critical. I think any good technologist
I've ever talked to knows they have a technical debt as soon as they write
the first line of code, 'cause maybe you picked
the wrong language, the wrong system, or you have the wrong infrastructure or whatever. So can we talk about this
idea of authentic founders for a minute? And so you told a story
about John and Logan and their story is incredible. They stood in the same, with me, I remember sitting in their office and talking about the clever of rides. So for you, does everyone
have to have an origin story, or how do you tell if they're, not everybody has a story
that's quite as good as theirs. And so how do you tell whether they should be in it, whether they're authentic? How does that work for you? - Well I really like
seeing some sort of tie. It doesn't need to be
ten years in the making. But I do like to see a tie
between the idea, and why you? Why is it that you have
that unique insight that no one else has? And I feel like for almost
every single investment that we've made, there is some sort of tie between that individual and
why they're doing that startup. But authenticity can come
in a lot of different forms. It could be that you are
actually the customer and so you have an unusual
empathy for the customer because you happen to be the
person you're building it for. And then the proof point
is, how big is that market? But the other could be,
I just see something. I've worked in this
industry for some time, and there's one team member
who really understands that market. And I always ask that question in every single pitch, and so I know that that's a important piece
of what we invest in. - [Professor] The question
you ask is, "Why you?" - Yeah, why you? And the reason why it's important is you know this, there's a moment in time with every single startup,
and it may be even two or three, where you march through the Valley of Death, and in that moment, if this was one of the ten startups you had whiteboarded on
the board in your office, and you knew of nine
other pretty good ideas, but this was just the
one that caught your eye in that moment, you are more likely to flee the scene in that first step through the Valley of Death,
and so we would rather have entrepreneurs who, it wasn't the tenth idea
that they came about. But this was the one and
it's pretty much the only one that they want to do. - Yeah. I think we go out of our way to say this over and over and over, Sam Altman did it the first class of CS23B,
and just say, look, starting up is kind of a crappy job. You can't, there's no room for anything else in your life hardly. And so you only really should do it if you can't not do it. So that's what I look for, people who are gonna do it irrespective of funders, irrespective of anyone. It just can't stay in. That's what Michael Dearing says, is that it kind of has to come out. - Yeah, it's like literally, you see it sort of physically
leaping out of their body. And so it was interesting
how you pitched your stack. And you call it a stack. And you talked about
technology and a product, because I think that a lot of people now are not doing the technology step. They start with product,
and I would actually kinda argue that Kevin
and Mikey did that, too, which was they figured out
product first, on Instagram and then they figured
out how to deliver it. You've got some companies
that are more like that, and some that are less like that. So how do you think about
technology led startups versus product let
startups, and is it really, do you really need both? - No. I don't think so, but
the proprietary level, we like to see some sort of insight there. - [Professor] Oh, so it's
not technology necessarily? - No, it's like, I would say
it's some sort of proprietary insight, whether it's supply
chain, or network effects, or it could be some fixed
cost that you're able to get, capitalize in some way. I think there's lots of
different things that can can kinda lock you in,
that's the customer, in a proprietary way. I think that piece is important. I would agree, most
things are not necessarily technology led. When we originally drew out the stack, I had technology at the
bottom, and then product. One of the things I would say was, some people just come in
at the product power level, and you can ignore the technology stack, and in fact, most of the
accelerator companies, those are ones that are coming
in at that product level. And so all you're focused on
is getting product market fit at that stage. - And then I guess the question is when you're looking at
technologies in IP as a driver... So you know, Gurjeet was one
of the first guys I met with when I started VNVC. This guy is unbelievable. He'll talk about, the
technology roughly is topologies, or topography? - Yeah, toplogy. - Topology, about how you basically make N dimensional pictures and
spaces of what's happening in your big data sets, and I was left with the observation that, man,
this guy's unbelievably smart, and he's built this
incredibly amazing thing, and I couldn't figure out
what the hell you do with it. And whether it would matter to
people on the product level. So how do you get comfort that this amazing technology
will bridge to something that is actually useful in the world? - Yeah, and I think the tricky piece, and this is, I invest in a sector, one part of my investments is
what I call radical science investing, and that's
Ph.D.s and some crazy piece of technology they
have, and the massive risk that we take when we
make those investments is that it's a technology
in search of a problem. And I think the reason why Ayasdi actually resonated for me was in my Ph.D., I actually had extraordinarily
complex data sets. They weren't big data in
the traditional sense. The files weren't that
big, but it was complex, and so it was hard to
see the actual important parts of that data set. And so I had some empathy for why being able to take very complex data sets and figure out what was inside
of it would be important. I think, so that actually
came more from a feeling. It was more of a gut instinct of I think there's lots
of data sets out there where you have this problem of it's just a big, hairy thing, and I still don't know how to query it. We also, at that same time period, we were creating, from
an investment standpoint, a thesis on big data, and one of the things that I had decided pretty early on was when
data becomes that large or that complex, I was
less interested in where you were gonna store it or
how you were gonna access it. I thought the big problem would be what do you do with it? And I had gone out and spoken to financial institutions,
some pharma companies to see what they were doing. And so from that perspective, I had some insight of, okay, these are the types of data sets
they're going to have, and probably, you might be able to use that in that setting. And so in some ways, that was lead by an investment thesis. We were looking for analytical companies. There's another company in our portfolio which is a company called Inscopix. It is literally a microscope you can put on the brain of a
mouse, and you can see the way it's thinking. And it was a Ph.D. student
in electrical engineering. And for probably five
years or four years now, we were sort of searching for what's the right business model? And it was one of these
things we don't really know how you turn it into this huge business. - [Professor] Direct to the mouse. - Yeah. - [Professor Direct channel to the mice? - Yeah, you literally plug
it into your USB port. Yeah, you could sell the mouse, like "This is what you're thinking." But what was really interesting was the guy was really capital efficient. So he took 1.5 million dollars from me and he never touched it,
and he sold something like 18.5 million dollars'
worth of this device. And he kept on iterating on, "Well I think the business
model might look like this. "I'm not so sure." And his advisers would come back and say, "Let's massage it this way and that." Finally, he's come back and said, "Here's what I think is
a billion dollar idea." And everyone said, "Of course. "Totally makes sense." And so he's off to the races, building a company around more like
a lumina for neuroscience, is what he believes. And so it's a big idea,
and he had the time to come up with that, because he was so capital efficient. So I think some of these will
take very different paths. Ideally, they don't spend a lot of money in that process, and I
think the places where we as VCs have struggled is, is it big R, little d, or is it little r, big D? And your hope is that you're skewed more towards development than research. - Yeah, I guess the lesson
for entrepreneurs then is you should figure out how
to find the right investor. Because I guarantee you, 90% of the people that Gurjeet pitched
didn't come at it from a, well my thesis was about big data, so maybe I'm looking
for, maybe I understand or have empathy for other
people who are doing it. That's interesting. I think the other thing is that, one other lesson here is that, assume nothing about investors as firms. Think about investors as people. Mike and Ann are both very different, just like Reid and I are very different. We have different backgrounds. If you do a little bit of research, you can find proclivities and interests, and it's not too hard. So you just learned a
little bit about Ann's patience around science. Now we'll talk about
marketplace in just a second, which also requires a
certain amount of patience. But then, and the funny
thing is, like one thing, you should be careful
not to assume too much, but you should check interests. So like everybody in the world assumes because of my background at Mozilla that I love investing in web things, and mostly, they make
me terrified right now, because of the rise of mobile. And actually, a lot of operators who, if you have operating experience, you mostly are terrified
of investing in that, because you know where
all the bad parts are. - Yeah, a lot of people... I mean, like, I actually
am interested in security. It was what was area of
research for me for my Ph.D. I haven't made many investments in this space because
it does still scare me. I feel like I know a little bit too much, but that's an area I
would still invest in. Another example is, a lot of people say, "Oh, she did optimizations, "so she's clearly gonna love ad tech." I don't. - Yeah, Reid went through a period where he couldn't really
look at a payments company who had echoes of PayPal. We would do it now. VCs always say that. We say, "We would totally
do it now, just not before." Okay, so let's talk about marketplaces, because marketplaces are special. It's a special category company. So let's talk about, 'cause
you have a few investments in marketplaces, so TaskRabbit
would be marketplace, the labor marketplace. Lyft would be in a marketplace. Modcloth is not, right. - No, it's more e-commerce. - Chloe and Isabell is-- - It's more e-commerce MLM. - MLM, okay. So, let's talk about marketplaces-- - Juanillo, might be. - Yeah, Juanillo kind of, yeah. So what we talk about is liquidity. How do you see, I mean
somebody asked about when do you see signs of
whether it's working or not. We talk a lot about are
the transactions clearing. Does that mean, do you
match a buyer and a seller, how do you tell? And then we talk about how different, difference if it's local or
physical or virtual or whatever, and we look a lot for
inflections in liquidity. How do you think about marketplaces? - Yeah, so at the very early stages, what I found is one of the most important pieces of a marketplace is
not demand, it's supply. And so at the very early
stages, we're looking at how effective are you at convincing supply to come on board, and how loyal are they to you as a platform, instead of trying to get a customer and then immediately move around the marketplace as a platform. And so we're looking
at the supply economics and how long term that supply wants to stay on that platform oftentimes, because that
actually leads to demand. So if you're really
great with your supply, and your supply loves
you and won't leave you, then that means that
your demand will come. And so you know, Lyft is
a great example of that. If you turn on your Lyft
app and there are no cars, the likelihood that you're
not going to turn on the app again suddenly goes up. And so if you look at sort of the way they do their marketing,
it's tied to how much the supply is well rounded in that area. I think that in TaskRabbit
we saw this as well. If you have a number of people bidding on a particular job, all of a sudden, to the demand side, it looks
like a really robust community and so people are more likely to stay. And so I like to see that
kind of supply dynamic starting to build in the early stages, 'cause that's the
leading indicator to what you're talking about
in terms of liquidity. Because once that happens, then
you can bring in that demand and you wanna be able to balance it out, but the balancing is a lot easier when you just have a lot of people who want to work in that platform. - Yeah, I think it probably depends on the type of company. I think what we talk about
is you gotta have a thesis about whether supply or demand matters. First, make sure you do
unnatural things to get that and then you're hopefully gonna
ratchet it up on both sides. And so let's talk about,
along those lines, let's talk about Lyft. Let's talk about how, can you talk about how the, actually, let's
come to Lyft in a second 'cause I think people are gonna be interested in in. Let's go back to the very beginning again. So what do you like to
see in founding teams? One of the things we asked Sam, is do you like solo
founders, team founders, two, three, 20. What do you like seeing in founding teams, and what qualities and what number? - Yeah, we tend to, and we've invested in solo founders, but I
think the healthier dynamic is to have at least two,
probably five might be too much, but two is usually a good number. Three is good as well. The reason why the solo
founder can be really hard is I found that it's very lonely. And as a result of that, it
just becomes this situation where you feel like you can't ever escape and you're sort of a
prisoner to this startup. And that dynamic can be really,
really bad in some ways. Whereas with a dual team, it's not just that you're a prisoner to your startup, you feel like this sort of social pressure to stay in the game. And so you can really sort of focus, both of them, or three
of them, on the problem, and it's sort of a teamwork element that I think can be really healthy. I also know that there's
no superhuman founder. Every person has a weakness. To the extent that you
can sort of round out one another's edges, it really can provide sort of different
perspectives into the company or the problems that you're facing. We like to see sort of that yin and yang. I look actually even within our own firm, and just the differences
between me and Mike, and you just look at us
and we look very different. Obviously, I'm a woman, he's a man. He's from Texas, I'm from Palo Alto. He's very white, I'm clearly Asian. Just our backgrounds, our perspectives can be very different. He's a lover, not a fighter. I'm more of a fighter. So I'll get mad about
something, and he'll be like, "Whoa, that's not a big deal. "What are you getting so testy about?" And he can really sort of bring me back whereas, he won't wanna
fight about something and so when we were negotiating our lease with our landlord, as an example, I get sent in as the pit bull and then he makes nice afterwards. So I like that kind of different roles for different people,
but not having all of it fall on your shoulders,
and I imagine for a startup it's like a thousand times worse. And so we're looking for that. We do like to look for someone fairly technical in a team as well, because we do think
that ends up developing a lot of that proprietary power. And so if you're a team
where you have a lot of business insights, and
you've entirely outsourced your development to Costa Rica, then we probably are less
likely to invest at that time. - Yep. So on the yin and yang about founders. There's a lot of different opinions. So a lot of people say,
well diversity early, and obviously diversity
in the long run's good. Keith Rabois or maybe Peter
Thiel is famous for saying diverse, you wanna
homogenate in the beginning so you can go fast an not
have to fight on everything. You guys don't worry about homogenating? I tend to think that homogenating and sort of types of
thinking around vision, around technology approach, that's kind of important. Is that not too important for you? - I think you have to have
an agreement on vision. So if you're arguing about
everything in the initial days, I think that's, you just lose time. So I think it's more about efficiency. So I think there are sort of fundamentals within a company that everyone has to be in agreement on, and
moving in one direction. That might be on technical stack, like we understand this
kind of technical stack. But I think diversity
in terms of perspective is actually pretty valuable. I don't think you can go from a homogenous organization to a diverse
organization later on. I don't think you can make up for that. I think it's very difficult to. - That's interesting. My theory is that I think
you can scale it over time. You introduce diversity, if you wait too long,
it becomes very hard, but I think if you start... It's interesting. It's an interesting question. You talked about hiring. You talked about hiring
as an important signal for you as to whether you
want to invest in somebody? When you invest in small, 'cause sometimes you invest in tiny companies,
couple, three people. How do you help them
figure out how to hire? Do you? Do you just let them... - A lot of times, it
depends on the company so some of our companies
will come back to us for hiring questions quite a bit, and we're involved in sort of even interviewing or closing candidates. The interesting piece,
though, that I found is they're asking us for candiates. That's usually problematic. So most of the time, the best companies are all finding people internally. So they're using their employee base to unlock the greatest
talent that is latent within another organization that's not even thinking of moving, that's not even thinking of a job. So our best companies are the ones where they're just picking
talent from great companies one at a time. They're like snipers. The ones that aren't doing as well, they're advertising, they're
putting up job postings everywhere, and they're trying
to found inbound interest. And so that's where we try to advise, who, within your team, who do they know? Because there's no one
who's gonna advocate for a company like someone
who's already working there and who's completely
bought into the vision. They're gonna be able to say, "I worked with this person two years ago, "and they were the
smartest person I ever met. "I'm gonna convince that
person to come work with us." Those are the best hires. And so we're trying to help our companies move in that direction, rather than, here's the recruiters that you use, and these are the job
boards that have been most effective. Those are sort of tactics
that you can use as well to augment, but
fundamentally, the best ones are internally found. - Yeah, I think that's my sense too, is that it's a binary bet, which is you're either a little bit past stuff and you're waiting for inbound people to come and talk to you, or you're saying "This is the problem I have to solve "that's critical to my business. "I'm just gonna go solve this thing." And figure out how to go
find the right people. And we have the same thing. I think it's easier in
technically deep companies to sniper shot. You can say, "I want the
best NLP search person." You can figure that out. So if you have other questions, go ahead and raise your hand as we go. I've got a couple more, and
then we'll just go from there. So let's talk about
Lyft for a little while, 'cause that's one of
the most famous pivots, well maybe Slack is more famous right now, but Lyft's pivot from where they had sort of carpooling to what's a little bit more like Uber became, and actually it was before Uber. I think most people don't know this. Uber X was really a copy of Lyft. And so can you talk about what the pivot felt like and how they knew it was time, and did it feel like running to something or what did it feel like? - So Zimride was one of these companies that was constantly experimenting. So they knew they didn't
have product market fit. What's interesting is they
had sold this platform to all the UCs. I think a hundred universities
had bought this platform. They'd sold it to Facebook, to Intuit, few corporations. - [Professor] It was sort
of like carpooling, right? - It was for carpooling. Lawrence Livermore Labs, and
so they had sort of this, one hypothesis was if
we could get all these campuses on board, and you could get geographical coverage
of the entire peninsula campus by campus. Then all of a sudden, you
might have something here, in interlinked networks. But that wasn't quite, it
wasn't moving fast enough. It was okay as a business,
it wasn't awesome. So they looked at bus routes, I think, from San Francisco to Tahoe and also van routes from San Francisco to LA. Literally, Logan was renting
a van and driving them between San Francisco and LA. We looked at bus routes
and enabling it for buses. I'm seeing actually a
lot of the resurgence of these original hypotheses that we had as actual individual companies now, which is interesting to see. But they had tried all of these things, but nothing felt like product market fit. You would book some people, you'd get some revenues. It was nice, but it didn't feel huge. And so Lyft was actually
just yet another experiment. They're like, "Well,
mobile's big, and maybe "if we do it P to P it's
sort of sharing the ride." I remember the original idea they pitched at least me was, "We're gonna have "women drivers and women riders." And by the time I got back to the office the next time and I had
been thinking about this, I'm like, "I think this is too small." They said, "Oh, we've already gotten rid "of that idea, and we're
gonna go more broadly." I think there were questions around how big an idea is this? How confident do you guys feel? And I remember there was this
real moment of intensity. Logan and John are really nice guys and so you don't get
that, I used to call it, "You need tiger's blood." They wouldn't have that intensity but they showed that
intensity in that moment and remember Logan was just like, "This is really gonna work. "You guys have no idea." And so the board said yeah, go for it. Let's do it. And I remember the
first week they launched we still had Zimride
alongside for a while. And the first week it launched, I wasn't really paying attention, but then Tommy Leap, who's a Stanford Tree, he came into my office, and he was like, "You have no idea how
huge this is gonna be. "I took four Lyfts in one day. "It was amazing." And I was like, "Seriously? "It was that good?" He was like, "It's amazing." And he became this super rider, and whenever we'd go to San Francisco, I'd park the car and
I'd take Lyfts around, and it was this kind of magical moment. But there was this whole Zimride platform hanging over their heads, and I remember this one walk where I walked with Logan around the Giants Park, and
we were walking, walking, and he's talking about,
"What should we do about this "additional asset? "Should we move people over onto Lyft, "it's really taking off. "How may of the people
should we move onto Lyft?" And at the end of that
walk, the conclusion was we need to move everyone onto Lyft. And now it would look like a no-brainer, but if you can imagine the
position that these guys are in, you're bought into a vision. When you pitch people and you've spent two, three years of your life, I think at one point before I invested, Logan said he was gonna eat from this huge can of beans for two weeks, because they didn't have that much money. You have sacrificed weekends. You have sacrificed birthday parties and time with friends and family to build this product,
and you're essentially coming face to face with the reality that this is not working,
and this other thing that you just built over a
couple months is really working. And you're actually going
to shut this thing down. That actually takes a lot of courage. It takes a lot of, it's this
huge come to Jesus moment that I think is really hard to appreciate, and I appreciate the courage that it took for them to move aggressively
in that direction. - It's actually hard even to see. 'Cause there's gonna
be a bunch of HBS case studies and stuff about
Lyft in the future, and what's gonna happen is say, well look, here's this thing that
was moving like this, and here's this other
thing moving like this. And so of course they shut
down the thing moving like this and they jump on the
thing moving like this, but we have so many cognitive biases-- - But it was causal. - It's causal and you
don't really wanna say, this thing that I've been working on, I should definitely look
for ways to shut this down. You actually don't,
it's hard to even see it because in your head, you're
so committed to the thing. - Yeah. And you've sold your
investors on this vision. You've sold yourself and
your employees this vision for a long time, and so I
think that when people say pivot and they've changed
the color on a button, it's not pivot. Pivot is actually, like
you're gonna throw up because of the idea
that you have to do it. - And so did they order a thousand pink mustaches right on that day? What happened? Do you guys remember this? All the cars had pink mustaches. - They had this outsource guy. He must have been like, "Oh my gosh, "I have won the lottery." - Pink mustache guy. - All the sudden they're getting all these pink mustache orders. I think eventually they bought
them out and he's in-house. - In-house pink mustache guy. How big do you have to
be before you need a guy in-house for your pink mustaches? - You wanted the cuddle 'stache, and then you wanted different
versions of the 'stache. - Did they run that by the board, or did cars just with pink mustaches just start showing up? - No, I think they mentioned it and I think everyone was like, "Oh, it's interesting." Why pink? - Sure. - [Student] How did the management team maintain employee faith
while going through this pretty uncommon transition? - Good question. - Well you have to remember-- - So the question was, how did the team maintain faith when
you're changing so much. - Yeah, so the entire team was fairly bought into all these experiments
that they were running, so they were used to, oh
here's another experiment that they're gonna run. I think the big shift was
they actually had a sales team that was selling this platform to universities and
corporations, and essentially you have to shutter that
and move everyone into, you're gonna operate this marketplace, which is, it's so different,
but at the same time, I think you get faith from
the fact that you have product market fit. They weren't just like, "Oh my gosh, "we need drivers. "Let's get as many drivers as we can." I think in the early days,
we had gone to TaskRabbit and said, oh you have some of
these people who are driving people to airports. Is it possible to use some of those people as drivers? And so, it was just sort of this shift that when you saw it taking off, I think everyone was bought in. - [Student] So how do you think, given kind of Uber's
rise in the marketplace, how do you think about
Lyft in the framework of all the things you
were talking about before? - Yeah, I think they have
a lot of the elements that we were talking about. One of the things that I've
always loved about Lyft-- - [Professor] The question was-- - Oh, sorry. - [Professor] What do
you do, so Uber seems-- - Uber, Uber's kind of-- - [Professor] Yeah, Uber, Uber. - It's interesting. We never talk about Uber. (laughter) I've always thought about
companies as being different, not better, and I think
that if you look at it, actually there are other competitors that did this P to P ride sharing stuff. And the reason why Lyft has maintained a great advantage here in certain pockets and markets, is because
they're really different. And one of the key
differences was when you first took a Lyft, in the early days especially, you had the fist bump, the pink mustache. It wasn't just everyone's private driver. It was such a differentiated experience, which is part of why Uber had to respond. But that community element has stayed, and if you talk to drivers,
if you talk to really long time users of Lyft, you'll censor what that community element feels like. I also think that sort of
the irreverence of the brand for a long time, has really created a differentiated appeal. I think it's partially why
I Lyft Line works so well. If you're in the city and you try it out, it's an incredible product. I've met incredible people
riding on a Lyft Line. I also thought that what
was interesting to me was, it's a experience that
translates generationally, so my dad actually had
a triple bypass surgery, and my mom and I both weren't available to take him to a doctor's
appointment at one point. And so I gave him this training on Lyft. I said, "Time comes, you press this button "and someone's gonna show up. "And you just get in the car. "You can sit in the front,
you can sit in the back, "whatever you feel
comfortable with, you go." And gets back and he
calls me, and he says, "What an incredible experience. "You press a button and
the nicest man showed "up at my door with a beautiful Prius "that was newer than mine
to take me to the hospital. "And he asked me about my life, "and if I was feeling
okay," which is like, not his experience he's ever had in any other setting,
and for someone who is 76 or 77 years old, to
feel that difference, to me, I feel like that's
something that's very defensible. I also believe that with them,
it's a marketplace model, and obviously there is a
huge battle for supply. And they all have their
different tactics for winning it and part of it has become
do you fixate locally? So do we fixate domestically in the US? Do you fixate internationally for growth? And Lyft has obviously picked their battle and how they're gonna go about doing that. And so it's different, it's
a different approach, and I think that's the thing that
we remain optimistic about. - Okay, I think we have
time for two more questions. It's you in the back. You can armwrestle. - [Student] When you are
investing in founders and companies, are you
investing in their knowledge about the specific market,
or are you investing more in the fact that
they're quick learners and that they're adaptive? Does that make sense? - The question was, when
you invest in founders, are you investing in the fact that they have deep knowledge or deep expertise, or the fact that they're
quick learners and can adapt. - [Student] Yeah, like
what if they just had the idea, but knew
nothing about the market? But you just think they have a good idea and think that they will learn. Which one, do you take
those factors into account? - Yeah, we would, but I think that, to some extent, we can
wait until another founder shows up who has both. And so, because we're not investing in a ton of different companies, we're actually fairly selective, so I'll probably make four
investments in one year. And so if you have one
very deep, and you're not, if I were to say which
one I would weight more, I would say someone who's
gonna be sort of fast. And quick learner means more that they're willing to
experiment and make mistakes and move on. But again, having some authentic insight into that market is better, and so we would prefer to see both. - What I'll tell you is that the market's efficient enough now. There's enough capital
and enough entrepreneurs that almost anything is
not an obvious decision. So what my experience has been, 'cause the pricing's
different or whatever it is, but my experience has been that when I'm warming up to an investment, I'll fall in love with
it, and then I will fall into deep hate with it, and
then I'll fall in love with it, then I'll fall way out of love. So I'll, like, this is
the best thing ever, like oh my god, if I invest in
this I might go to jail, and you vacillate wildly,
and then at the end it comes, and you make a call, and
then you usually throw up after you make the call,
and it sounds ridiculous for an investor to say, but it's nerve wracking on our side, too. And so I would just say
you vacillate wildly and it's a mixture of things. And obviously the more
good things you have, the better, but everything's got issues. - Yeah, I think the key
thing is most of these things are controversial. So in the early days, if
it's not controversial, there's so many people already working on that very same idea, that you probably don't have the space to win. And so there's something
about the business, as I said, it's less than
50% chance of working, and so, and you'll go
through these wild swings of like, "Of course I believe in it. "It's amazing. "It's gonna totally work." And then you're like, "Oh my gosh, "I'm such an idiot. "Why do I wanna write a check?" And then you go back to, "But I love it. "This is totally gonna be amazing." And you have these wild swings. If you didn't have that, and especially within our partnerships,
it's just two voices, and oftentimes, the best ones are the ones where Mike is like, "I don't get it. "I don't see it at all. "That's the stupidest thing I've heard." And I'm like, "Well,
I'm gonna do it anyway." And the opposite has happened as well, where he says, I'm pounding
the table, and I'm like that is so ridiculous, and
I'll literally make fun of the idea, and he says,
"Still wanna do it." Those are the best ones. - [Professor] Airbnb,
Facebook were both very, very contentious in our partnership. They seem obvious now. Last question. - I said no, we passed on Airbnb. - [Professor] Yeah, well it's like you're renting out somebody's room. It's so weird. - Well they had more money off of cereal than they did from... - [Professor] Yeah, last question. - [Student] So product market fit is the gate to category dominance? But to what extent be
concerned that some things might have a great product market fit in Silicon Valley, in
California, in the United States, but not to a global audience,
and consumer software is kind of dependent on a
global audience at some point. - Yeah. I would say, actually for
consumer, US audience-- - Question is how much do you worry about product market fit
is local or geographically based and not more broad. - So yeah. Geographically based, we, if it's local we really worry. So one of the things that
Mike and I often talk about is, is this a blue state
wine sipper business or does it have a chance of being a beer slammer red state business? And we prefer businesses where we can say that's red state, beer slammer business, because too often, we'll
find businesses that cater to the triathlete,
VC, who loves to drink wine than, that's great, but that's
really not gonna translate outside of a few people. So that's something that we
do very much pay attention to. International versus US. I say we focus pretty much on the US. If you can really claim the US market, you already have a fairly
large, substantial business. - [Professor] Okay, thank you. Thank you very much, Ann. - Thank you. (applause)