- As a quick intro, Reed Hastings is widely
known within the Valley as being one of the most
principled, purposeful, rigorous in his analysis and intelligence about being a founder and being a CEO. He gets requests all the
time to join people's boards, including mine eight years
ago, or whatever that was, which he was like, "Hey,
I'm focused on education. "Go away." Which is a part of focus, is one of the things that
makes Reed unusually strong. So, we are extraordinarily
lucky to have him with us here today. And I think what I will
open with is another piece of something where you are famous for, which is the Netflix culture
and the Culture Deck. How did the first Culture
Deck end up coming about, getting published? There's version two, which we'll get to, but what led to that
creation of that deck, the publication of it? What was that as an act of
management, thought leadership? - It wasn't a pre-read, so there's no pressure on it, but how many people have already read it? Netflix Culture Deck. Okay, that's fine. Thank you. So, about eight years ago, we were getting tired of
new employees joining, and I would meet with
all the new employees and go through this 100-page slide deck. And two out of three really
got it and understood, and their managers had properly
described to them Netflix, and one out of three were in shock, because the deck had some
aggressive statements like, "Adequate performance gets a generous severance package," and "We're a team, not a family." And so, if you know that going in, you can love it. But if you don't know it and you didn't expect it, you feel bait-and-switched. And so the big driver was we realized every candidate should get it, and that, of course, if every
candidate's gonna get it, it's gonna be public, so let's just make it public. - Was the construction of
it a way to essentially, as you're doing scale to
kind of keep a cultural norm as you scaled? Was that one of the central
tools for doing that? - Yes. I wouldn't quite say keep a cultural norm, but it was to share and debate. One of the great things about
putting things in writing is that it makes it more debatable, and so it's not just hearing
a sermon and it's like, "Okay, that's nice," but you can comment on, "You know, this part here that we wrote down doesn't
really seem appropriate, or doesn't seem what we do." To some degree it's a bill
of rights for employees, which is "Here's a set of
things that we, the company, want to and aspire to operate by, and if we don't, you can call us on it." And then we'll either fix
the articulation because it's too easy to misunderstand, or live up to what we want to. So, in some ways it's aspirational, and then by putting it in writing, which probably started
five years before that, it allows all the people in the company to collaborate on it and
to suggest improvements. - And what other things do you do to keep a healthy culture at scale? What are the other kind of techniques in addition to the deck, in addition of the on-boarding? - Yeah I would say in our
first couple of years, so we're founded about
20 years ago in '97, to when we went public in 2002, all we were concerned with
is not going bankrupt, because the first big
internet recession was 2001, and so we had built up a big expense rate contingent on future financing. NASDAQ fell 75% that year, and all the venture capitals
felt over-leveraged. They were barely gonna be able to support their existing investments. Certainly no one wanted
to take on new ones. The year before, in 2000, to raise money was as easy as taking a tin can and shaking it and like $50,000,000
would show up in the can. It was incredible. I've never seen anything
like it until last year. And then the following
year after the crash, you take your tin can out and shake it, and someone would steal the can. So, it was an enormously
sharp and short change, which is why you see so much
talk today about bubble. No one cares about the bubble, they care about the fall. So, suddenly we didn't have enough money, we had to lay off a third
of the company in 2001, and we eked into profitability and survived. But until that point, we really spent all the time just on how are we gonna get the business to be, we were DVD by mail at the time. So, it was, how is it gonna be profitable? And only after 2002 we realized, "Wow, we're gonna survive." Then we started to think, "Wow, it would be awful to not want to work here, so we should try to figure
out what cultural attributes were there of what we wanted to work in." We didn't set out to try to say, "What's the most theoretically
efficient culture?" in some abstract sense, it was literally about us, and we realized that what we valued most, even more than success, was working with really talented
people in a productive way. That that joy of
excellence was incredible. And of course if you can get
that joy of people excellence and talent density, then you're very likely to win. So, they come together, but at the core it was
about having dense talent. And part of it was, after we did this one third layoff, it was 120 people to 80, we had expected to
basically grind to a halt and really not be able
to make any improvements, cause it would take 80 people
just to keep the lights on. But, in fact, we got more
done with only 80 people, and we tried to figure out why. And we realized now there was
no dummy-proofing necessary, and so it was just
everybody was going fast and everything was right. And so, we realized with
the right density of talent, there's very little process needed, and that that was the joyful thing. So, at first we said, "Let's do a one third layoff every year." That that was the key. - That's a long-term winning strategy. - That's a long-term winning, it might be. Back to the blitzscaling thing. And then we realized. - Shrinking, we're gonna now add. - Blitz shrinking. Well, you want your gross heads to be 100% and just your net adds are less. But talent density matters so much. So, then we figured, "No, let's not do a layoff
every year like that. Let's just do continued focus." And the big evolution we
came to is for managers, they have to each year testify
for each of their people that if that person were trying to quit, they would try to change their mind. And so, it's up to the manager to decide, "Yes, I would want to change their mind," but sometimes you find that
you have people working for you and if they quit you'd be like, "That's pretty good." And that's the case that
we wanted to proactively, not wait for them to quit, but proactively give them a
generous severance package. - Yeah, and that's how I
come to the Culture Deck. In addition to performance, cause this is, actually I'm to some degree, leading the witness on this
or leading the question, but what were the key things
about establishing that culture which enabled you to scale? Because this is the highly-aligned, loosely-coupled part of your
version two of the deck. - So, I think when we went public we were 150 people by then, and people worried internally. Now that we're public, everything's gonna go to shit, because so many companies had done that: put in a lot of process,
be very cautious, etc. And we've made great progress really every year of pushing back on that and doing something
either symbolic or real that increased employee freedom. Essentially incented more
variance in what we're doing, and we realized if we're
gonna run really loose, like very little rules, then you do want to set a broad context. So, we set context, and this is what Reid's referring to, which is we added essentially
a chapter to our Slide Deck, which is "Context, Not Control." And so, we try to get managers to inspire and lead people rather
than micromanage them, and they inspire and lead them through setting the context. "What are we trying to do? What are the constraints? Is it a really big problem
or a little problem? Must be done right, or we
do an approximate version and fix later?" There's a lot of context
to any problem area. And if you get good at setting context, than you don't have to
direct the micro-specifics. But it's an art. So, there's context about the problem, and then there's context about behavior, which is really the culture. Which is, "what are the behavioral norms in terms of honesty and
sharing and forthrightness?" - And in terms of setting the context, one of the things when I read, we've actually never talked about the version two of the deck, and when I read that and
compared and contrast and looked at it, I was like, "Alright well one of the key
things is people usually want command and control." And you have an exemptions
slide on the control, but people usually wanna
be able to do something, like a bold new innovation
usually requires somebody going, "This is my idea, I'm driving it home." How does setting the context
and that kind of innovation, how do those line up and
what do you do about that? - Well, I think you might be mixing, the manager shares a lot of context, but a lot of people make decisions. We definitely have decision owners. There's almost nothing by consensus, but you gotta know the context and then you make decisions and a lot of things happen
that I don't even know about, about different deals or
different things or hiring. You guys are gonna get lucky and meet with Marisa in a week or two, and one of the things that saddened me when she went to Yahoo, which was very brave of her to do, was this idea that she was
gonna review every resumé to raise the bar on hiring. And to me that's a suicidal methodology for a senior executive because you got a thousand people and you're so tactically buried trying to do a thousand resumés. And I think what she was going for, and you can ask her, is a symbol of, "I care
about the details." A better way, there are
better ways to do symbols. I don't review resumés,
I don't block hires. I look at it as my job is
to evangelize the benefits of super-great talent density. But we don't even try to hire perfectly, because when you are interviewing, it's six or eight hours with someone. It's something, but you know them so much better after they've been in for three months, so I'm like, "If you have an instinct
you wanna try, try." Take the person in. If it doesn't work out, we'll give them the severance package and we'll move on and it's not the end of the world. So, we're experimental
with people in the same way you might be experimental
with a product feature or something. And it's important to be honest with a person when you're doing that. But we don't look at it as a marriage. There's no need to, and that's essentially an
adaptation to the growth. - Well, actually I think
what Marissa's trying to do a little bit, because we had Eric in here a couple of weeks ago. Three weeks ago. Anyway and part of what
Eric was expressing about how they built
the Google culture was, "You guys should only hire
people with certain kinds of degrees from certain
kinds of Universities," where they were actually, in fact, reviewing every resumé, every CD before. I think she's like, "Okay, we gotta get to that kind of setting of the bar." And that actually kind of
leads naturally to one of the. - But that's mistaken in so many ways. First of all, it's trying
to do it the Google way, which even, anyway. - Well, that's precisely the
question I was getting to, which is there are other attempts at building strong cultures. Google is one of them. How would you kind of compare and contrast so people understand that there isn't just Netflix is the only strong culture, but here is some way of understanding the variances between strong cultures? - Well, I think the key thing, and it's a great point, is there are many different cultures, and as long as you're really
clear about your culture, you can get the employees
who love that culture. And not everyone will be the same fit. There are some things, like integrity, that are pretty universal. But there are other things
of how to deal with dissent, where there's perfectly
reasonable different ways and effective different
ways of dealing with things. So, the literature on
culture is pretty clear that strong cultures work, and weak cultures are diffused. So, in a weak culture, it's really a diverse culture. You get many people
acting very differently, and then they don't understand each other and they feel undercut and political. Even in a, let's take a Hollywood studio that many of you would
not like to work in, they have a very clear
system of how to operate in terms of how they manage the politics, and it's internally consistent. And everybody knows it and the people who get good at it love it, and a lot of people don't. It's not inherently awful or ineffective, it just wasn't right, say, for us. So, that's why I think
culture is an expression of what you and your
senior team want to be. It's a bit aspirational in that way. - Is part of the interviewing process, do you have essentially a culture check? - Yes, but not like a checklist. Culture is always something
that we interview for of kind of curiosity. It's easy to find people who say, "Oh, I've read the Netflix Culture Deck. I love it, I really wanna be there." What's harder is, "What are the main things you disagree with and why?" When we get a blank stare, we're like, "Okay, not really a
first principle thinker." And when they say, "Well, I thought the way that you didn't talk
about how to acculturate," and "if I'm not great
on day three, am I out or do you look at it like an athlete where it's over some time
period to prove yourself? And if so, what's that time period, and why haven't you clarified that?" Then, we're like "Oh, yeah
that's a good insight." First principle thinker. So, we're looking for
people who are curious, typically self-confident, and they're questioning
everything around them. - Last culture question before we move on to some of the other kind of interesting lessons from Netflix, missions. So, I think you can only
have a strong culture when you actually have a
mission in the company. How do you define the Netflix mission, and how is that kind of
embodied in the strategy, in the work-play? - Well, I would say I've always believed that a strong mission
statement is really important, and I have to say we've
never succeeded at it. It's like this odd thing. We've had them every year for 15 years, and they're constantly different, and now we're all so cynical. We're like, "Come on, we're
trying to make people happy." I mean, that's the fundamental
thing in consumer service. We're trying to be better, and us and Disney and mission
aren't that different. There's a bunch of
different tactics at it, but we're an emotional product. We cater to the things that are
important but not necessary, and I love that part about it, cause when you're selling
milk or penicillin, you're selling something somebody needs and we get to sell something you want that you choose to do
because it's really human to be entertained and connect
to people over entertainment. And so, we celebrate that a lot and we want to change the industry, but in terms of a mission statement like so many companies have, Facebook talks about a more
connected and open world, and then you see that phrase
again and again and again, and it's a powerful mantra. We don't have an equivalent and I think we probably
would've been stronger if we do. In the early days it was to connect people with movies they'll love. And it's okay, but it's kind of like a long, kind of awkward phrase, "Connect people with movies they'll love." It's just not punchy. We used it, we had t-shirts. And then it didn't really motivate anyone. It didn't really stick. But we knew we were trying to make people happy with DVD's. And then, eventually it was
streaming and original content. So, it didn't hold us back much. - So, let's go on to some
of the lessons from Netflix. So, if you were to call a
younger version of yourself and tell yourself to do, what key lessons that you would do differently through Netflix, through the history of Netflix. Obviously, you've been a founder of companies before Rational et cetera. But what were the key lessons you would, "This is an inflection thing that I wish I had done differently?" - We went public in 2002, growing nicely in '03 and '04, and Blockbuster which then
was 20 times larger than us, finally counter-attacked and in a. - I'm just curious, how
many people, raise your hand if you know what Blockbuster is. - Oh, they know what it is (laughs). - It's not quite. - [Reed] A more accurate question was, "Do you remember actually renting a DVD? And Redbox is still quite popular. In '03 and '04, we're growing really fast. '05 Blockbuster attacks. And so we thought we were very clever. We came up with a number
of ways to counter-attack, and remember that
everybody's got every DVD, so Blockbuster and us had
the same content offering, cause it's non-exclusive
licensing to get DVDs. So we could differentiate. On our service, fulfillment
levels were 99%, and theirs were 89. But that's a really
abstract for a customer, thinking of signing up
doesn't mean anything, right? And then they discounted massively, so it's like half our price. And so, we were losing share. So we were struggling with what to do about it. So we did three or four big efforts. One is in the end of '05
we added Netflix Friends, which was our own social network. Remember in '05 Facebook's
just on a few universities. So our own social network
amongst our members, and if you enabled each other, you could see what each other
were renting and rating. And we thought, "Wow, the viral effects of
this would be really powerful," et cetera. We added ad sales. Back in the day it was
like Yahoo type banners, and we sold banners above the choosing interface. We added used DVD sales
on our site to consumers. Of course we had some excess
DVDs from four years ago, so we added an operation to
sell them for four bucks a pop. It's a different logistics,
but we added all that. And we added Red Envelope Entertainment, which was a group that we brought in to buy films out of Sundance
and similar festivals to then publish them on DVDs, so getting into content. So we did these four efforts. And each one was a
dozen people, 15 people, and made us feel great. Employees loved them because here was a tangible thing that was not just 98% versus 89%. Here was the thing that they
couldn't do or weren't doing, and it was a differentiator, right? And wasn't management so clever? We went through these waves
of battles in '06 and '07, and in the end we won. They ended up closing
down their online thing, and two years later going
completely bankrupt. But we looked back and
none of those four efforts made any contribution to our victory. And one by one, we had closed
them down along the way, every single one of those four. And so, in hindsight, we
realized when attacked, we should retreat to do the core better and not try to broaden the
surface of attack, essentially. And it was a great lesson for us of focus, So now when people say, "Aren't you getting into news or sports?" we're like, "Absolutely not," and we're really confident in our answer. Movies and T.V. shows on a global basis, enormous market. And so we're much less
subject to being prone to go off and chase the shiny object to try to have something in
the checklist to differentiate, and to trust. And in hindsight if we had just gone from 98% perfect to 99.9%, we would've done a lot
more for the business. And it's hard work. It's operational logistics,
how to get DVDs not to break, the amount of polycarbonate
analysis we had to do, it was all this stuff. We would've beaten Blockbuster
sooner than we did. And so that's one of the great
lessons out of that time. - What about a Qwikster? - Yeah, Qwikster we got
some lessons out of. So, I'll give you some context. Now that we're growing, '07, '08, '09, we're adding streaming,
it's a hybrid plan, you've got DVDs and streaming. '010 we launch in Canada
as a test, in a way, of both international and streaming only, cause there's no DVDs. And we expected x and we
got 10x bigger reaction and customer signups than we expected. I mean, it was just massive even though it was just
streaming, no DVDs, and we realized, "Oh my God. "You don't need DVDs." - Wasn't that part of the
plan from the beginning? - Well, we were always gonna
use them as a hybrid way. This was our differentiator against Hulu, because Hulu was about
equal size at the time. "But we have DVDs too." That's our differentiator. "We're streaming and DVDs." Which was gonna be compelling, was compelling to half the market, but was gonna be compelling
to a shrinking percentage of the market every year. So, it's not what you want
your brand image to be around, so we realized, "Wow, we cannot afford to use the crutch of DVD now because pure streaming works and so we wanna win as the best
place to do pure streaming, not be the best place to do hybrid, because that wasn't a long-term benefit." So, we had to not rely on DVD. So, I think correctly, we
separated the two businesses. The huge mistake we made, because remember, in 2010 we're in the middle of a
recession still pretty deep, as we said, "Well, $8 is a good
price for the DVD plan, it's a good price for the streaming plan, so the combination, which
is $16, is the new price." And the old price was 10 bucks. So from the customer standpoint, one day the price went up 60% in the middle of a recession, and now they had to deal with two user interfaces instead of one. So, this was not well regarded, and if we had done
grandfathering on the price for a couple of years
it would've been fine. Now, fortunately we realized, "Wow, we've got to make amends." But now a lot of people have left us and changing the price would be hard, but the symbol of Qwikster was much hated even though the name
tested quite positively before we did this, it had attracted a lot of negative. So, if we killed the symbol, maybe people would forgive us. And so that's why we killed Qwikster. And people mostly forgave
us and we moved on. - And one of the things
I learned from that, from you when you were doing that, is I actually thought that
your letter to the public on the blog was actually masterfully done. It was personal, it was open. Was that just a natural
reflex as part of leadership to write that kind of letter and make it as an open statement about, "Look, we need to over-communicate, we need to be humble," or was that something
that kind of was a result of a ton of internal discussion about how to do crisis management? - Well, I think everyone
who has both Oxford and Stanford degrees loved that letter. - Oh. (audience laughs) So, I loved that letter. - Yes. No, in hindsight it actually was a very bone-headed move for your typical consumer (laughs) because it was. - New lesson. - It was an explanation
of basically why we needed to do this thing and
they didn't really care. It was about our survival,
blah, blah, blah. So there was a few more people
than just you that liked it, but by and large it was tone-deaf consumer marketing, because they didn't hate the letter, it had some positives of being authentic, but it just stirred things up. And a better thing would've just been to change the situation
i.e. Qwikster or whatever, as opposed to get into these. It's just too intellectual. - So, if streaming wasn't
part of the initial plan, what was the initial plan? Was it just DVDs? - No, streaming was. In '87 I took Tannenbaum's
networking class here, and I don't know if they
still use the same textbook or some adaptation of it, but you have to calculate the
bandwidth of a station wagon filled with backup tapes
driving across the country. And of course it turns out
to be terabits per second and it's got a high latency, three days, to get any bit to the east coast. But it's super efficient
digital distribution network. And in '97 when someone showed me DVD, it was the first time I had seen it, I was like, "Oh my God,
that's the station wagon. That's five gigabytes on
this one ounce plastic disc, and I'll bet you can mail that." And there's my network,
there's my station wagon. And so, we always conceived of DVD by mail as a digital distribution
network with high latency. And someday we'd use the
internet to take the latency out. And early internet was more expensive to deliver five gigs than mail. It only took another three
of four years of Moore's Law to make that cross over. But we always conceived
of DVD as temporary, internet streaming as the long-term. - Was there anything in the initial plan that turned out to be radically wrong that your engagement
with the market changed? - Yeah in '97 when we were
first seeking funding, we thought it would be mostly
streaming in five years, by 2002. And in 2002, there was nothing. It was still dial-up. AOL was, the company was in trouble, but dial-up was still king. And so, around then said
by within five years, streaming will be half our business. And by 2007 it was still zero. So we said once more the next five years it's gonna be half our business. And we were wrong but the other way. The next five years it was like 60%. So we just kept repeating the same thing, and eventually it turned out right. - Yeah, one of the things that I sometimes like to say about predictions is the only prediction about the future that is generally true
is that most predictions of the future are false, but it's the paradoxes. A number of people here
are actually entrepreneurs, some of them are applying
at certain companies. Is there anything in the financing process for early Netflix, lessons you learned that you would redo like in terms of how
you ran the financing? - More relevant, I think. I graduated in '88. Worked at a research lab, worked at a networking company, worked at a startup that failed where I was just a hardcore engineer. And then I had an idea. And you really want to do
the entrepreneurial thing when you have the right idea. There's no point of deciding, "I'm gonna do it this year
with the best idea I've got." You're probably gonna have crap results. And so it's when some idea
is just shaking you so hard that you'll go into poverty
to see that idea materialized. And the idea is really important to you. And then you change your life
to become an entrepreneur to do the idea. And so the concept or whatever
it is that you have has to, and it's good to constantly
be thinking about what could that concept be, but I would approach it that way. So, my first company was funded. 20 people, 20k a piece. So broader family and
friends and angel round. Only one or two semi-professional. No one close to Sam Altman
or someone like that. It was a little less
developed 20 years ago in the ecosystem. On that money, we were
able to launch a product, get some initial traction, and then we got our first venture capital, which was the predecessor to
Benchmark, Merill Pickard, and Mayfield. And that worked out, I
would say, very well. We got people that we
trusted and really helped us grow the business. So I had an enormously positive
venture capital experience in that way. And what you realize is
many third party vendors and banks and things, they will screw with your company because they don't really care. But they care a lot about their reputation with your board members. And your board members, when they're venture capitalists, become the guarantor that
you won't get screwed with. At least, much less anyway. Because their long-term reputation is with your board members,
not with you, typically. And so that's a big, subtle value add. In Netflix's case, I can self-fund, I had done well enough
on the first company which got sold in '97 that I self-funded the first couple of rounds. So it's not relevant, probably,
for a typical situation. In hindsight, we wish
we had done more debt and less equity, and you
know the cardinal things. - The prediction from success. - Yes, that's right. Nothing too dramatic. - How about your first hires? How would you think about, cause you've now done this twice, both times successfully. Once massively successfully. How do you think about the
first five to ten people? How do you think about doing that hiring? How do you run it? - It's people you wanna work with, but you can't be afraid to change. You can't let friendship get in the way of professional judgement. So, you've always gotta be thinking, "Do I have the right team?" and, "I owe it to the whole
team, to the marketplace, to my investors, to myself, to do what I think is the
right thing for the company." And when I was first a CEO, I'd never managed anyone, so I'd never written a review
for anyone or anything, because I was an individual engineer. I found a lot of those
communication things difficult, and in particular I couldn't
possible fire someone. It felt so cruel, it felt inhuman. It's like breaking up with
a girlfriend or boyfriend or something, but worse. It felt awful because it was selfish. So, it took me probably three years before I came to a view that I had. And the rest of the company
was depending on me, to do the right thing. And that it wasn't about me being selfish and pushing the person out, it was about me protecting the company. And once I could neutralize
the value scales that way, I could comfortably say to someone, "I'm sorry for this, but I
think it's best for the company if we end your employment." And once it was, "And I think
it's best for the company," I could get there. Now, many of you are
probably more sophisticated, more mature, you've broken up
with more partners (laughter) so it's easier for you, cause it just takes practice. We don't get enough practice in that. - Actually, there's a book
I recommend to people, which I don't know if you've encountered, by Fred Kofman called
"Conscious Business." And I'll get you a copy, but
it's actually very similar which is thinking about compassion and compassion management. The compassion is not
only to the individual, the compassion is to the company. The compassion is to the customers. The compassion to the shareholders. You actually have to
integrate all that compassion, and once you're doing that,
that actually makes the, "Oh, yeah I understand
this is painful for you, but actually, in fact, it's
still a compassionate act." And that was actually
one of the things I found really good in how Fred
thinks about these things. - Yeah, and I think one part of compassion is a severance package, because when you cut someone loose and it's a week or two of pay, it really can be difficult for them, but our minimum is four months and it goes up from there. So, even if two weeks in we realize there's a mistake, as long as the person hasn't
lied to us or something but we've made a mistake, they get a four month severance package. And that does a couple of things. One, it makes the person
feel better about it cause they know it's a
material amount of money. And again, four months is the minimum. And two, we get a legal
release in exchange for that. And so despite the fact that we've separated from
a thousand people, maybe, we've never had an employee lawsuit, because of this. And then third and most subtly, managers are really nice people. They're human beings and
they're good with people, so they don't like doing mean things. And so, you can think
of the severance payment as a bribe to the manager
to do the right thing, because it makes it easier. And without that, what they do is they put an employee on a
performance improvement plan, and then the two of them go
through this excruciating dance where it takes two or
three or four months, and then they break up but
it's clearly documented. And that's just a lot worse. So, that's why we avoid that. And essentially we buy out and then the person's good. They'll take a month or two off, and then they'll get snapped up somewhere. - When did you start with the
four month severance package? - It was three months
from the very beginning, - That's unusual for a startup. - Yes. It's real money, but it's worth it, because again, otherwise the person, it's costing you anyway. It doesn't really cost you any more, because you're going to
spend the three months managing the person out, and so you're gonna end up with that hit. It's kind of a classic. - All right, so let's... actually, no. One thing. Has your hiring changed as you scaled? Like your hiring practices? Cause you're very familiar
with all blitzscaling pieces of how all these things change
at each different magnitude. Was there anything notable
when you look back at hiring that you said, "Oh, these
are things we added in as we got to larger scales?" - No, I would say we have a very big people judgement process. Some companies try to standardize and put a lot of
statistics and analysis in, and we've never believed in that. We say, about data, use a lot of data when you're picking stocks, but probably don't use a lot of data in picking a spouse. That the more that you're
in an emotional element and a gut feel element, the less useful data is. And the more you want to really listen to your intuition and judgement. So, we spent a lot of time with candidates trying to get them to talk and feel, but then we don't do like a how-long-did-it-take analysis, or I don't know. Something like that. And then, we do a lot on references. I'm always stunned how many companies don't do good references and how many employees that
I don't get called about. And I'm like, "Really? "You didn't even call?" And so, we're manic on, and always blind. I mean, we never even ask
the person for references. There's this great tool,
LinkedIn, (laughter) and so you can find the reference. That has changed the game in references. - So, let's go to some of the areas where Netflix has done
some very interesting, and actually unique strategies that, as far as I can tell no
one else ever did them, and there may be additional ones to the ones I'm about to ask you about. Bu let's start with the prize for the recommendation algorithm. Where you kind of said, "Okay, this is one of the most
important parts of our business. Which other movies would you also see? How do we essentially
crowdsource and get network intelligence for the best possible ideas. Let's do a prize." How was that conceived? How did it work? Would you do it again? Would you do it differently? - I'll tell you the story of it, but unforunately it's disappointing because it turns out that we did one, it was epic, and we've never
been able to do another. So, it's not like a general purpose thing that you can use a lot. Once in a while, there's a problem that is
amenable to outside analysis. That one was. We hit the right tones on it, it was transformative for
Netflix in the machine-learning community in 2007, and really helped move things up a notch. And still today we get residual goodwill from it. And so of course we want to
do prize two and prize three and prize four, and we still have not found even a single other one problem that was amenable to that kind
of formalized architecture. So, very specialty, exotic
tool as you grow companies. - And say a little bit
about how you did the tool. Is there anything you would've changed? I understand it's rare. And then, what you got out of it. - I didn't say it was an open contest. We put out a data set, and it's
like, "Crack this data set. to this, reducing the root mean squared error by 10%." And you could submit these attempts, because you're doing new
algorithms like in that, but you can think of it like
cracking a cryptographic thing, or any of those like pay for
bug bounties is a good one that people do as an example of a crowdsource one. But again, there are these little pockets where it's a workable solution. - Did you end up getting
a better algorithm, or did you end up getting
mostly better goodwill? - Much bigger benefit in goodwill. Some benefit, the algorithm
worked and we put it in place, but then we grew past it. So, it wasn't like a
permanent step function, it was a one-time. - So, now let's go to original content. This is probably one of the things that's probably most stunned most classic Silicon Valley people because
we think we're good at tech. That's what the talent base here is. We're bad at content programming, that's what you go down to L.A. for. It's the kind of thing
where intensely engineering and technology driven culture will be intrinsically bad at this, because we'll be bad at people, we'll be bad at emotions. And so when you start doing this, I think the general kind
of "received wisdom" in the Valley was, "Oh, that's a mistake." And then, of course, you pull it out, you do it much more successfully than Amazon, also doing it. And then later a following. And then other folks. What led to the strategy? Why is it that you just literally trounced conventional wisdom on this? What were the lessons from that? - I say Amazon is doing
a pretty reasonable job but we can come to that. Necessity is a mother of invention. Every cable network. FX, AMC, HBO, started on other people's content and then got some scale in distribution. And then went into doing
their own original content. So the path for Netflix
was well understood like other cable networks. So it wasn't like, "Oh my God we never thought about this" kind of thing. You get more differentiation and love and value from if you did the content well. Again we tried it early
in 2005 on a DVD basis. We only had five million subscribers. And the economics were roughly for a movie like we got those cool Maggie
Gyllenhaal movie at Sundance. We would have normally bought, suppose Sony Classics had bought the film. We would have bought 20,000 DVD's at $20. So $400,000. But instead now we can buy
the film for $2 million and put $2 million to marketing it. So we had ten times the cost. And we still didn't need
more than 20,000 DVD's. We're like, "Wow this sucks." (crowd laughs) Ten times as much! And meanwhile Sony
bought some other movie, made it popular enough with our members, and we still had to buy the 20,000. So we didn't really save anything right? And we realized, "We're just not nearly at scale." And we got out of it. And then in 2010, we were spending enough
on the big-ticket TV shows from a licensing standpoint that my partner Ted Sarandos, who lives in L.A. said, "Now is the time." And we were like, "Oh my God are you crazy?" We're gonna put a $100
million that was like a quarter of the content budget. And he was like, "Don't worry I've negotiated. We don't have to pay for it till 2013." (laughs) - Don't worry the bill
will come to you later. - The bill will come to you later. It's due on arrival. I'm like, "Oh yeah I feel great now." Man it was transformative. So he picked incredibly for what we need and I don't get any credit for that. That was all him. And now Ted's built a team from studios that are cherry pickings. Some of the best people out of CBS, NBC, to do exactly that. And they really know how to run shows. And the main thing is
the cultural attributes that we talked about before, context not control, freedom and responsibility. Those have played extremely
well in our relations with the show runners, with the producers. And so it's been a very consistent model. And it turns out that really
our culture is designed at inspiring and
motivating creative people and you can create a software person. Create a marketer. Create a content producer. It's really in the creative world. And the creative world the hallmark is, "You do your best thinking in the shower." You have a huge idea. The day after Thanksgiving
on a mountain bike ride trying to recover. It's not like you need eight, ten, 12, 15 hours in the office
doing something because in the creative community it's synthesizing all of these things. And that's true in a number of fields. - And is this, was the cost reason the reason you did? Cause I think all of
your original content's been TV shows not movies. Or is it most? - Most. We just had a big one, "Beasts of No Nation" was our first big movie. And then we have one with Brad Pitt next year called "War Machine." - And was the reason
because of the cost of TV? The reason you focused on TV? Or was there a different
reason to initially focus on TV and then build there? - The economics of TV are much better and there's too much capital
chasing movies because of it's historic legacy. So the returns and the movie business are just much harder to make money. Yes, TV's less expensive
per hour typically, but not for any inherent reason. Just cause that's what other people spend. So to be as good as them, that's all you have to spend. So it's really influx of capital. And then of course when you
fall in love with the show, then there's a built-in
market for season two. So that's a great thing. - And how is the original
content playing now? What are the current things that are knocking out of the park? What have you learned from them? - Huge hit right now, "Master of None" anybody seen it? A couple, the rest of you
have got to go see it. I'm telling you. Watch the first two or three episodes. Aziz Ansari and it's
his commentary on life. And it's unbelievably funny. And he casts his parents in it and gets away with it. And it's both warm and delicious, and biting and hugely original. So that's our current one. "Narcos" was a big one this summer. Which was American company
us contracting with Gaumont which is the oldest film producing company in the world. It's a French company. Filming in Bogota, Colombia
with Brazilian talent, and it's super popular in Germany. (crowd laughs) - It's defined globalism. - That's right, and so this
is why Netflix is racing to get as global as quickly as possible. We want to produce
content all over the world to share with the world. And so it's not a Hollywood to the world, it's really connecting great stories. - Is there anything
that when you look back at Netflix you think
was a kind of a strategy that you pioneered that we didn't cover, that would be useful for the students? - Once you're public, which would be a great day, you're gonna have an issue
around stockholder utility and stock comp options and RSU's. That really frustrates people because most of our company stocks
are pretty volatile. And then if the stock goes down, people feel cheated. So I think the mechanism, if you get to that stage, look up how Netflix does
stock options because it's from my experience in
the first company Purisoft with dealing with that
volatility that we did a big adaptation in Netflix. So I think works much better. And it's basically price averaging if you follow stock markets. - So let's switch to a little bit of the, kind of the classic role of a job. Of the role of a CEO. How do you define the role of the CEO? - That definitely varies
by stage and company. In the first couple of
years you do everything. You're doing dishes at night. You're coding, maybe you're writing
marketing materials for sure, you're dealing with
customers and investors, and you have so many
disadvantages as an irrelevant little nothing as a
company that you have to make up for it with talent and hard work and brute force. And so if you're lucky, that's only a couple of years as opposed to ten years where you're just on call constantly in a very intense way. And then the trick is as you get to 50, 100 people you've got to evolve your management style as you get to 500 people. Kind of each 10x, maybe even just five x you've got to adapt to be more strategic because it used to be you knew every person and now you don't. So now you gotta manage
a little bit differently and still you wanna be a great leader for those people. When you get to real scale, most of what I do is
vision about what markets we should be in. Like we should go global, we should do originals. But I'm not picking countries and I'm not picking shows. Vision of we should spend 5%
or 10% of revenue on marketing as opposed to 50% or 2%. So some big macro decisions like that. But not what's the campaign. So there's vision there. There's vision in terms of culture. What are the rules of the road of how the firm operates? What's our character? So that it's a healthy place. Vision, focus, inspiration. But you really, you can't do much of the work because it's just too big. So if you tried to, you're laboring yourself out and getting everyone else upset. In my first company it was
probably a more relevant example for all of you. I was 33. The company's growing about 50 people, and I was still trying to coat it. We had 50 people. I was still writing product at night, and then trying to be CEO at day and sleeping at work. And I wasn't careful enough
about taking showers. And finally someone said to me, "Shower!", (crowd laughs) and it's just gross. We don't want our leader to be gross. We get that you're working hard. And by the way, when there's bugs in your coat, it's really hard. - It's a different definition of culture. - It's really hard to get
you to fix things because you're off doing other things. And so I was trying to hold on too long to the dual roles. - And so how do you improve
your ability as CEO? How do you learn? How do you be on a constant improvement? - In my first company, I think I made a mistake and I felt like investing
in me was sort of selfish. I should be working. And so I got invited to YPO which is Young Presidents Organization, and it's a good learning type paradigm. And I looked at it a little, but I was like, "It's like a day a month. I can't take a day a month off of work." And for me it was like off of work. And in hindsight that was a
real mistake cause I didn't. I was too busy chopping
wood to sharpen the axe. As of later state CEO, you need the skills and personal maturity and growth and reflection and I should have cross-fertilized. I should have spent more time
with other entrepreneurs, seeing how other people do it, learn and invest in myself. Maybe that's do more yoga or do meditation or be more. It's like that great Gandhi quote about "Oh my God I've got so much to do. I've got to meditate for at
least two hours this morning." I didn't have that perspective. Really until that company sold, and then I did a wonderful thing that Reid Hoffman also did
being a crown fellow. I didn't understand that
connection that by making myself better both as a person and a leader, I was actually helping the company even though I was away from work. And it sounds so obvious to say it, but I just didn't give
myself permission to do that. - Is there anything you
encounter cause one of the things you also have a general
reputation for is a deep certainty of conviction of belief. Have the things that you
encountered as CEO caused you to question yourself? - I would say that's not
an accurate attribution. It may come across that way, but I'm very much like Hank Paulson and you can never know anything for sure. Maintaining a reservoir of
doubt is really important. We often try to do the
exercise like at Netflix. You're 18 months with doing exercise, what would be different in Netflix if you were CEO? And it's a way for me to
gather all the things. And some people say, "They've changed the pricing. If we do this different we
would be in the ad business." And then I really try to think through. A lot of "Would I really do that? Am I sure? How confident am I?" We call it farming for dissent. You're constantly, you can never be confident of anything. You want to constantly be curious, and yet you have to be
executing really firmly so it's a little bit like your lovely Isaiah Berlin piece of
simultaneously knowing that there is no certainty and yet working really
hard to what you think. - It's Isaiah Berlin's
two concepts of freedom. Anyway for. - For another class. - For another class. - You should hear Reid's discussion. I guess he's waiting till
the end for the questions. - We have about 25 minutes. Then we'll go to questions. - Also partially, this is in part a
portrait to how you think so the people can then go and ask specific questions and the things that they know that you will actually be amazing at. So board of directors. How do you compose them? Is there such a thing
as a high performance board of directors? And what would you do
differently with your years of entrepreneurial
experience doing this? - When you are interviewing
venture capitalists, you definitely wanna pick on the people as opposed to the evaluation. The quality of the
people in your board room is super important to you. I say quality. It is probably not the right word. The rapport you have and the connection you have and your trust makes a huge difference. If the people in that very, that's like a marriage. It's very hard to get divorced
from your board members. You gotta spend some time with them and then say, "Okay if they're in, they're in." And then you gotta listen to them, and then you definitely
have to be honest with them. You'd be amazed how many CEO's are not very honest with their board members. And that's poison because then the investor has
to constantly be pushing and probing and trying to
figure out what reality is cause that's their job. So you want to really lay
everything on the table and be the paradigm of honesty and straightforwardness
with your board members. And if you've got the right ones, they won't freak out. They'll be very thoughtful. So I would say there's
no perfect board members. It's unique to each of you of someone whose judgement you trust. Whose integrity you trust, and you enjoy being with. - Do you think the role of the board changes as the organization changes size? And if so how? - Yeah when you're small
or your board members know a lot more than you as a company cause they pattern match on
a lot of different things, the larger you get, the more any board member. I was on the board of Microsoft
for a number of years, and on Facebook you
spend two days a quarter on a public board. And you're just level of depth. There's not near as strong as the CEO's and the senior leadership teams. You're more of a safety net then. Then you are able to contribute. - And then the last in terms of the roles, and then we'll get to the
blitzscaling part of this. But how should someone
think about knowing whether or not they're a capable founder or not? How do you kind of say "I am. I should actually do this", versus "I shouldn't do this"? - It's a good question and I'm not sure I didn't know. I just cared about the product of this first product idea I had. I wanted to see that out. And I was definitely reluctant CEO. A couple of times I
screwed up badly enough that I said to the board, "Let's get someone in who
knows what they're doing and I'll be VP of engineering and I'll be fine." And both times they said, "No you did screw up, but we still rather have
you than someone else and the dislocation that that brings." So that's why I say I was reluctant. And the whole first company
which was '91 and '97, I was kind of miserable. I was always underwater. The company was always, as soon as I learned something, the company was bigger. I was killing myself physically and mentally trying so hard and just underwater. Then that got acquired in '97, and we started Netflix. That's like 15 people that was like compared to the other company had 700. It was like so easy. So then Netflix has been fantastic and total joy for me because it's kind of like your
second kid or something where your first kid it's a panic. "Oh my God." And depending on the kids, as long as they're both healthy, your second kid is much easier and you're much less freaked out. So for me then I have enjoyed the last 18, 19 years tremendously because I've been able to keep
up with the challenges. - So let's shift to
scaling and blitz scaling. One of the things that you and I have actually had
a conversation about for as like sometimes these things
are just normal scale up and sometimes there's blitz scaling. Netflix is actually more of an example of kind of just regular scaling. There's relatively little
blitz scaling within it. But it's one of the things where, actually in fact the consistency, the expecting growth also matters and being systematic
about how you change size. What's your sense of the key characteristics to know in advance in order to scale right? And then what's your
kind of reflections on the kind of the different levels? - Well let's see. There are scale economies in businesses like Amazon and Netflix. The bigger you get, the cheaper the marginal customer sale is. And those are powerful. Those merit you investing
forward to get to scale, and so running losses for
the first couple of years doing things like that. So call that normal business. And then there's a couple
of rare businesses like LinkedIn and Facebook that
are network effect businesses. And then the prize of getting big first is so much larger than just a scale business. Then in those businesses, it's worth selling 90% of the company to raise a $1 billion, to be the big one. You must be the winner. It's such extreme winner-take-all. I think it's the degree of winner-take-all and network effect that gets you to either take the risks and the stress of blitz
scaling rather than the more. Like Netflix, like Amazon, in the early years we're
going 80% year over year, and then 70% and 60% an hour 25. And 25 at our size or
Amazon's are big numbers. But that's like a normal
great scaling business. And both Amazon against Walmart and us against cable and satellite. We have huge markets. Growing at these rates is great, and hopefully we'll be able to grow for 20 years. And Amazon the same. But then there's a couple of businesses. Again eBay was one of the first where it's real network effect. And in those businesses, I think you do kind of
crazy practices because you've got to grow 300% to maximize the opportunity. So you're gonna be sloppy
on a number of things. So you probably don't
worry about talent density and you don't do a bunch
of the things that I described which are the
things that are appropriate if you are trying to be a
careful disciplined operator. But if it's "get big or go home", and then that's the interesting thing. But it, again, most
businesses is not that. It's businesses that have network effect. - Well some of the things that
I think interesting about it is sometimes it's network effect. Sometimes you have to
get to a minimum scale, which is Amazon and you. And there's a number of other businesses. The thing I think is, would be fun to get some
perspective on here for you is when do you think that it's necessary to move fast because of competition? So like for example, well another thing I
think drives scaling fast is you say, "Okay I've got to occupy the
green field market before them" cause presumably this is a
little bit of how you think about global now in terms of, it's not just a question
of content in globalism, but it's a question of, "Look we've proven this thing and let's, in every relevant
market in the world, this is something we should be there." So no? - No, or yes and no. A company can get started
two or three years before us in Germany, and we can knock them out in 12 months because we have great scale, and it's not a network of facts. Where with eBay tried to
knock out Yahoo! Japan, they didn't succeed and that's been permanent. So if you're in the
network effect business, then you get more of first is forever. Cause the barriers are really strong. But if you're just in the scale business, you can go in with a lot of content and energy depending
on what your market is and that's why Amazon. It's like how quickly
did they wipe out CDNow in like 1999? Like really quickly right? Because it didn't matter that
CDNow had started earlier, more scale on that side. Again I think it's really
related to barriers and these network effect
barriers are so strong. - Although just as you're saying, scale barriers are relevant too. Cause once you get to a certain size, you can leverage your
size in order to win. Cause that's precisely your Germany point. - There are scale economies. They're just not as strong as a barrier. So you can think a barrier
as how much pricing power do you have, which could either be how much advertising or what the literal price is. And if you've got a ton of pricing power, then you can raise prices a lot and it's still very hard
to come after you because you've got that network. - Were there any, as you got larger in
terms of the organization, were there any new
strategies that you added? Cause a classic one is
corporate development in terms of that's something
you now do when you're hundreds of people versus
earlier or in the thousands. Were there any kind of
key strategies that got, obviously original content
was the one that came out and scaled? - Yeah not particularly for us. So in 17 years we have
never done an acquisition. Obviously we haven't been acquired. And corporate development is M&A. Again we haven't had those
kind of dynamics in our market. There's other markets where
the key thing is to buy up a bunch of companies
that might be competitors or adjoining to get to
scale and differentiation. - So let's do a last few questions
on kind of Silicon Valley and then open up to the group. So what do you think makes
Silicon Valley so unique in the world? - Well there have been a
lot of companies like Amazon that are scale businesses, not network effect that are successful outside
of Silicon Valley cause you don't have to be here I
think to do a scale business. So far all the network effect
businesses are here because it's this network effect is when you need to do blitz scaling. You need to be in Silicon Valley, or it's like ten times better
as opposed to only 50% better. - So it's here or China. - In their markets yes. Fair enough. So China's such a self-contained market. It's an isolated island because of the great firewall that it just, it might be a different planet. It's got its own ecosystem. - But huge island. (crowd laughing) - Okay, it's an island for the rest of the world that
headquarters has been here. - And why do you think Silicon Valley tends to be network effects? - I think it's historic factors that get Detroit to be the
center of the car world, and then there's more talent
there to build car companies. You get those kind of. You get Manhattan and London for financial products. So you get density. It's almost irrelevant why it started. it's just that once it happens, then it's the best place to be. And then continues to be a better. I mean it's a much better place now than it was 20 years
ago for entrepreneurs. And then that was better
than 20 years before that. The more of you that do companies, the more there's that ecosytem. And so our benefits as
an area continue to go. Again but if you've got like a minor idea, that will be a $10 million
business in ten years, pretty good really. Move to Austin. Do not do that here because you're competing with blow-away ideas for people and talent and money and cost of talent. - Are there any myths that
you think we tell ourselves that Silicon Valley people
that are worthy of exploding? - Not obviously but. - I'll give you one
that I think is kind of, actually in fact I think a
lot of Silicon Valley people think that they're actually
very good at managing and actually in fact, they're not bad at managing but actually in fact it's the
composition of how you put together like network effects businesses. How do you do a disruption
in the industry. Those kind of things
create a lot of things. It's not actually in fact
genius management techniques. - Do you think Larry Page
or Steve Jobs would've said they're good at managing? I think Larry Page is amazing scientist, and he's a really deep thinker, and he's an amazing guy to inspire and lead Google. But visionary in so many ways. But not like you would go into GE and be that successful. He doesn't think of himself
as a general manager. And so I don't think most
Silicon Valley people. Tim Cook might cause he grew up on that. In operations and he is
a very talented manager. So it just goes to show that, you look at how successful
Tim Cook's been at Apple. It's been phenomenal and he's obviously a very
different person than Steve was, and different management style. Which just says that there's
multiple styles that work. I think you shouldn't
try to be someone else. Don't like read about Steve Jobs and then start wearing
mock black turtlenecks. (crowd laughing) Be yourself. Oh God here I am in black. (crowd laughing) Figure out what your emotional center is and be that really great. And don't try to be other people. I mean it's fine to read about them, but there's no one style that. And if you look at the varieties of the really successful companies, the variety of leaders is quite wide. - [Voiceover] Can you talk
a little bit about the switch from DVD's to streaming and how you internally
organize the product teams? Did you buy everyone and bring the different teams or did they change their duties? - So question is how do
we evolve the internal management to move from DVD to streaming? Because the first
streaming product offering was integrated into DVD. If you were a DVD subscriber for Netflix, you then got a little
bit of free streaming and then you got a little
more free streaming. It was pretty integrated as opposed to it was a separate offering. So it only got really
difficult when we got to a separate offering. There's a streaming service from Netflix, and there's a DVD service. And the only really painful part was probably 2010 or 2011 when out of 30 VP's we had five focused on the DVD. And this was a close knit exec team and then we kicked the
five out of the meeting. And we realized we have to eat, sleep and breathe streaming like Hulu. And not also have this great five people. And of course they felt bad, and eventually they realized
they're great at DVD's and they should follow that. So there was a time of separation. But in the first part it was integrated, but that's because the product
offering was integrated. - In the back. - [Voiceover] I'm curious about how the regulatory risks has
helped with your strategy as to play this technicality and international expansion. Has this affected the
core of business now or? - Well as you grow. - [Reid] Repeat the question. - Yes thanks the question
is how a regulatory strategy has been important. We've had to fight for net neutrality and a big threat for us
was that cable companies who are most broadband would get to charge us a tremendous amount to be
able to serve our customers as opposed to be pretty open and that they would have
an advantage cause they wouldn't charge themselves. So we've had to fight that
regulatory battle over the last four years. And then every company that gets large faces anti-trust concerns over time. You see that with Google
in Europe right now. - [Voiceover] Why don't you police users sharing their
account with other users? That must be something really
deep that concerns you guys. - So question is why don't we
police or crack down on users who share with other users? So when a wife shares with
her husband the password, it's probably okay. When the dad shares with the kids that are six years old living at home, it's probably okay. When that kid goes to college, it's probably okay. Remember 90% of kids live at home when they go to college. So that's definitely okay. So now the kids hanging out
in the apartment next door then maybe that's okay. Or is it not? Or when the kid arrives at college, is that okay? So the only thing that's really not okay is the dorm room. So the dorm gets an account. The fraternity gets an account. And 30 people do it. And there's an easy check
on that which is you only get one or two concurrent
streams at the same time. You're not gonna share too broadly. And to get more concurrent streams, it's more cost. And so that is an effective
limiter on broadscale sharing and then we don't have
to try to police people's living situations. It is three friends? Three friends in an apartment, is that legitimate or not? We get to skid out of all of that. - [Voiceover] So I read on Netflix. On the website it talks
about sports mentality when it comes to managing. I'm just curious what's the benefit and some like setting my account back? - Sure the classic, it's probably not for you guys but it used to be a classic
metaphor for companies was the family. And you'd hire them and you'd be like, "We're a family. We look after each other." And it's kind of baloney because you'll lay someone off in a way that you wouldn't your sister. You want them to work for you
like they were a family member for free and cheap, but you're not really
prepared to treat them like a family member if you're honest. Really the professional
relationship is like a sports team. And if you want to win a championship, you got to have incredible
talent at every position. We said, "Look we're like a professional sport", not like your kid's soccer team, but a "professional sports team where we pay people well. We want them to win." And if you have one bad game, you don't lose your position. But ultimately you're fighting
for your position every year. And that's how we feel about it. As long as we're honest about it, it's exciting cause then you can play really sophisticated sports. So to do a blind pass in soccer and you just know the person's there, that's an art. You need great teammates. That you are so well rehearsed. To do that you need great talent and that's fun to be around. - So we'll get to Chris's
question in a second, but that's too good of a layup. Comments on Chris and
my's book "The Alliance". - Meaning in short terms of duty? You're signing up to get something done, which is like a player's contract. We don't tend to use
contracts in Silicon Valley, but Hollywood does. Their practice is formalized. You sign up for three years or five years and they mostly stick. And so the next evolution
of the alliance is introducing contracts
in the Silicon Valley. Again we're not used to it
from a labor standpoint, but large sectors of the
economy work that way, and it works fine. It's a different system. - So the first part is, can you talk about a
few of the other things that you decided to do
differently as Pure Software? And then the second part is
you've talked about investing yourself like what are the
things that have been most productive in terms of
return on investment and investment itself? - Sure. The big thing that we
do differently as Pure we were really trying, which was '91 and '97 to be efficient. We were engineers running a company, and so every time something went wrong, we put a process in
place to make sure that didn't go wrong again. Which is a very semiconductor
yield frame of reference. We saw everything as a
magic set of processes. But then the market shifted on us and in that case from C++ to Java, and it turned out that
everyone that the company had were well adapted
to following the rules and implementing the processes. But that doesn't make them the
people who are well adapted to re-think the market and how should we approach
this other market. The company struggled mightily and ultimately was sold because of that. What we realized is we were over-optimized for one business, and we had led the company become, it's like monoculture agriculture, we had one strain of
wheat over a huge area. Very efficient in the short term. And then what we should have
optimized for is flexibility. And that in fact all
companies face challenges when market shifts so we
should have been more tolerant of inefficiency and mistake and error, and more focused on flexibility because I think in the long term, flexibility is what's chosen for. So if you think of human evolution, we're not very well adapted
to many specific things and we humans are super
well adapted to be flexible. And so our lifeform has
conquered the world because of that flexibility
despite many maladaptations for any particular climate. Of course there's many
examples if you think about it. And the trick is management
is generally short term. You're focused on making
these better next year, and that's mostly efficiency seeking. So you have to balance
that with some long term about the value of seeking flexibility. - [Reid] And then the second part was? - Best thing for, it probably varies by person. For me it's been reading combination of western civ and leadership and management books. "Beyond entrepreneurship"
is one of my favorite, Jim Collins before he did "Good to Great" and "Built to Last." This book in '95 or '96 he published it. And the first 80 pages you can read that every year for ten years. It's really perfect for the entrepreneur, "Beyond Entrepreneurship." - [Voiceover] Can you talk
more specifically about how you incorporated the machine learning and the AI side of
things with your own data set up Netflix, and the wild crazy world
of Hollywood in producing? How do you evaluate
content that's passed on, especially now since you're
gonna have a lot of pitches? - It is, we basically,
it's a good insight. The question is how do
you apply machine learning to picking content or developing content? We basically don't. Your intuition as how
would that be relevant. What shows up when you
open Netflix on your iPhone or on the web or on the TV
that is a ranking problem. The ecosystem is pretty
good at ranking problems. So it's a big ranking problem. What are the titles most relevant and interesting to you? So it's really big on that aspect. Now we can apply that on marketing also with real-time bidding and personalized marketing. But in terms of picking shows, that's like picking employees. It's a human judgement task
of what's gonna be big. Once a show's produced, we can selectively promote
it on statistical techniques based on who's enjoying it, who's watching it really quickly versus watching it slowly. What other people are like that? So promoting the content to the
right subset of the audience which is critical when you've
got a big global audience, is a very machine learning problem. But the actual production, like other people have tried
to do script analysis systems. When you take a script and you code it up in all these dimensions and then try to do predictions. And none of it's worked. So there's so much in
the artistic execution that you as a consumer
then love from the script. And once it's produced, there's no value in predicting it. You're just gonna put it out and try. - [Voiceover] So as a quick bog, so is the focus of Netflix gonna be one show for larger audiences or smaller shows for niche audiences? - The question is are
we focused on big shows or small shows, and we're agnostic. We're focused on intensity. We want you to love what you watch, and sometimes that will
be a big show cause you're in the mainstream of
the society that you live in and sometimes it will
be a neat show because you're not at the mainsteam. But again it's all about intensity. We have to match cost of production to the expected size of the audience, but we're just as happy to
do small shows as big shows especially as you get to
personalized marketing externally and of course you've got
personalized merchandising within the service. - [Voiceover] So it seems
as though you guys get a lot more creative direction to the people that you obviously contract to make it famous. I was just wondering how
you balance that with making sure that you ensure
these are good shows? - So the question is how do
we ensure these are good shows and how much creative direction do we do. It depends upon the show. Generally we like to meet with a team and then if we like them and back them, let them run. We're very hands off once we've made a decision to back them. If they want in pub, we give it to them. We'll do these things. But relative to other studios, we're much more letting them run, and we've won a reputation for that in the creative community. - So last question from
that side in the jacket. - [Voiceover] So especially
when you're first starting out, how quickly should you go to your board and your board of advisors, and how do you think about
the limited characteristics? - You know it's a much
better ecosystem for that, especially the board of advisors level now than it was 20 years ago. I ended up not having
any board of advisors and it wasn't that common. I think Reid Hoffman
would be a much better much more insightful for you on that. Sort of the subtlety of the
current startup ecosystem or Sam Altman that you've met and just not that qualified for it. - So with that let's
thank Reed for joining us for a lovely time. (audience claps)