Importance of Natural Resources

Blitzscaling 09: Reid Hoffman and Allen Blue on Why and How They Scaled LinkedIn


– So we’re gonna start the same way we did with the last lecture, I’m
gonna move in the center, and basically we wanted
to sort of reflect on, this is our schedule for today, so we’re gonna do the same
thing we did last time. Start with a recap and themes from OS2 and then we’re gonna move into OS3. Talk about some specific areas of OS3 and then talk a little
bit about what’s coming up on Thursday. We’re gonna start, though, by talking about what we did at the beginning of OS2, which was what did we hear
from the three speakers who came in and the three
practitioners who came in and what are common themes that
we actually heard from them? Just reflect on that as a bit as a group. Anybody got common themes
we heard from those folks? – [Voiceover] Control communication
as the company scales. – So you have to control communication as the company scales up. Communication is a different problem as the company gets larger. What else? Yeah. – [Voiceover] Hiring
is absolutely critical when scaling your business. – Yeah, yeah, hiring is
absolutely critical scale in the business. – [Voiceover] Metrics. – Metrics, say a bit more. – [Voiceover] If you don’t
have them, you’re screwed. – (laughs) If you don’t have
metrics, you’re screwed. True and everybody talked about metrics and how important they were. It’s interesting, there’s a set of things which had to do with measuring wind and measuring your progress forward. Metrics still tend, in the OS2 level, to be a fairly lightweight process, where you’re not having
to put a lot of hardware behind it yet. And actually you begin to
see that, really, in OS3. What else? – [Voiceover] Maintaining a
small group sense of leadership and growing equal from
within your (mumbles). – Yeah, so maintaining a small group and growing leaders from
within if you possibly can. One of the things that’s just
gonna be really interesting is at the OS2 range, when
you’re still a small group of people who basically,
sort of a band of friends or brother or sisters, you
basically have the ability to stay close as a group
and to grow from within, but when you’re actually
in a hyper growth phase and you’re moving especially to OS3, the requirements for executive leadership can be quite onerous and very unexpected for the group of people
who are highly affiliated and ready to work in startups. So we’ll talk about that a bit today. Anybody else? Any other things you
thought that stood out from that group of people? Like things which they disagreed on or even disagreed with the
first three speakers on? – [Voiceover] I thought
Eric’s hiring practices were pretty counterintuitive
to the rest of the valleys. – Hm, so Eric’s hiring
practice is counterintuitive. Say a bit more. – [Voiceover] That they
only take top people from top colleges and disregard the rest. – (laughs) Right. Agreed. Yeah, seriously, I mean Google
definitely has a reputation if you want to comment on that, ’cause it is different than
the usual hiring process. Not that that isn’t a great
direction for hiring folks. – Well, the key thing when you’re thinking about establishing a culture
and actually having a distinct, it’s actually, in fact, not all A players will fit in the same culture, so you’re actually
trying to find something where even prospective A
players won’t fit within it. So one of the things that was particular to the Google culture is they said, okay, what we are about
is this elite university, generally speaking, tech education and that’s what we are. That’s part of what the
centerpiece of it is. So it helped them preserve their culture as they were getting
larger, but it also means that you have a particular,
it’s not like any time we can find an A player
whatsoever we hire them. It’s actually fit within a culture and most cultures have some,
not necessarily exactly like an elite degree, but
they have some version of this is who we are and
sometimes you run across an A player that you go,
“Oh, you’re an A player, “but doesn’t fit the culture.” And in preserving culture
as part of what’s important is you begin to get to
questions of communication, questions of leaderships,
questions of making a more decentralized set of decisions. If you don’t have a unified
culture, you start breaking. And that’s why a unified
culture’s important, and that’s part of the reason
why it’s counterintuitive, but it is one of the ways to establish a kind of firm culture early and that was one of the reasons, well, I don’t know if
they did it explicitly for that reason, but that
was one of the benefits they got out of it. – [Voiceover] I was gonna
say, are you gonna talk about what LinkedIn’s version of that was or should I ask that now? – [Allen] Yeah, we are
gonna talk about it. – Yeah, that’s in a few– – We’re sorry, we will talk about what LinkedIn’s version of
it is just for the recording. Okay, so I’m gonna hand
things over to Reid to talk about OS3. – Okay, great. Let’s see here. Right. So, the first two weeks are
essentially the startup class. What we’ve done the last,
four weeks actually, sorry, first two sections are like,
how do you have an idea, how do you think about it,
how do you validate it, what are some of the
considerations for that. Part of the reason why
we, like, for example, when we open with our guest lectures on these topics are people who see a lot of early-stage things
in order what to see. And when you get to OS2 and
you get to, essentially, the tribe, you might be
moving the scaling quickly, but a lot of what you tend to be doing is essentially figuring it out, making sure you have product-market fit, sometimes you know already. And it’s really when you get to OS3, e.g., hundreds of employees and roughly speaking, if you
wanted to be technical about it in terms of the way a lot
of Silicon Valley people look at this, it’s kind of 150. So when you say, is there
an exact numerical bound, there really isn’t. It’s a kind of tenor of an organization. It’s how it’s working, but
you’d say it’s about 150 when you shift from 10s to 100s and that’s what we’re calling the village. And that’s actually in
fact when you start getting to the really substantive shift and this is, to our very first class, this is the call out and what happens is when you begin moving from
the tribe to the village, you’re beginning to say okay,
’cause there’s this thing called Dunbar’s number,
you can look it up, it’s on Wikipedia, but roughly speaking, it’s up to about 150
people all know each other. You have perfect hyper cube,
everybody knows everybody else. When you move past that, it
gets actually very difficult to do that and so coordination
becomes more challenging and you move past it,
not because you want to, but because you need to. And usually you’re beginning
to get into some real traction, whether it’s on consumer
or business scale. You may or may not yet have
that kind of revenue stream, but you’re beginning to see
what your revenue stream is. It’s not a formula across it. And that begins to get
you into the question of when you’re scaling from tribe to village, are you scaling from tribe
to village in essentially kind of one of three ways. One way is, look, we just naturally need to get a little larger, we
don’t know how much larger we’re gonna need to do,
we’re gonna scale up some, we’re gonna solidify the organization, and we’re gonna make that organization essentially a
quasi-indefinite organization, which is we’re gonna run with it for a while and we’re gonna see. That’s the least scaling, that’s the least of what we’re talking about
in terms of blitzscaling. The second one is, look, we’re
gonna be scaling, call it, and that’s probably, by the way, you’re scaling 20%, 25% organization, etc. We’re gonna be really scaling
and we’re gonna be doing, somewhere between a 15 and
100% organizational growth. All right, 100% being double, but we also know that the likelihood is we’re gonna continue past it, that we’re just gonna keep doubling. We’re gonna maybe go to tripling. We’re gonna add in on a regular basis and so when you’re building
out your organization, you’re not building out an organization that you say, okay, this is stable, ’cause you realize when you double again, you’re gonna change your
organization structure, you’re gonna change how you communicate, you’re gonna change what
your leadership looks like, you’re gonna change
how much multithreading you’re doing in the organization. You’re gonna make all of these changes, which are really key. Now, you’re not doing it at
kind of a totally hyperspeed, but you’re doing it at a thorough speed by which you’re keeping
your organization flexible as opposed to hardening
it the way you would do in the first instance. And then the third choice
is, oh my gosh, right? And by the way, not all the
companies have to do this, but, oh my gosh, I have
to more than double. I have to triple, I have to quadruple. That was the kind of thing
that Eric was talking about last Thursday, which was the,
look, doubling’s pretty easy, (laughs) right? Quadrupling is super hard. And this is where you fundamentally start making that choice. The earlier places, very, very rarely are you actually in
fact facing that choice. Like, in PayPal, when we were doing it, we essentially got to about 80 people and we realized we had to do it, so we actually realized the tribe, but that was because it was like, you very rarely have a
circumstance where you’re getting this compounding number
of transactions everyday and it’s compounding
on an exponential curve and then you know you just need to go. Usually you’ve started figuring it out, now you’re pushing for scale. And so, to recap, part of
the problem you’re trying to solve, oh, this is the
family versus the household, but yeah (laughs). We’re calling it family versus household, we were wrestling with the term. You’re trying to say,
okay, do I have something that’s a unique idea? And do I assemble a few people,
do I try to make it happen? I may raise a little bit of money, but I’m trying to figure that out. When you get to the tribe,
you’re trying to say, okay, do I have enough sense
of the product-market fit that I can actually put on the juice that I should really double down, that I should try to occupy that market. You may be doing a lot of
pivoting in this circumstance, maybe doing pivoting later too, depending on how it plays out, but you’re really, generally speaking, trying to figure out product-market fit. And then when you get to the village, what you’re trying to do is say, okay, now that I have a sense of
what this can be globally, how do I scale it up? And there will be a couple
of different questions, like one question will
be your market size, one question will be,
do I grow into something that’s worth taking a big shot at it? Second question will be capital because you can’t actually,
in fact, scale fast without some real access to capital and that real access to
capital can be, for example, in Google’s case, can be revenue, right? That’s a great place to be, where you’re essentially
printing so much money, that you’re like, oh,
we got as much capital to do whatever we wanna do. It’s, by the way, a hard
problem too, but it’s ideal. But what you see a lot
of modern variants of, you know, you see Uber
doing, Airbnb doing, Dropbox doing, a bunch of other folks, you see them raising a bunch of money because if the story works,
if they’ve got enough of the early positioning
of the team in place, they’ve demonstrated
the product-market fit, they can demonstrate a growth curve, then actually, in fact, raising
money and using that capital to scale in the thing
fast is one of the things that most typically happens here in Silicon Valley and China. And this is the place where you, almost certainly making that
decision in order to go. That’s part of the reason
why the village is so key and so you have to say, okay, we got here. We’re gonna scale more. Frequently in the very valuable
cases we have to do it fast. Now, part of the decisioning
on doing it fast is, what does the global
competition look like? That’s pretty central. And so, for example, when you look at some of the differences,
part of the reason that Miriam was actually, in fact, kind of not trying to
hyperscale what she was doing, but taking it in a very
considered way was because her competition was old school
market stationers, right? No one in the new realm
was actually, in fact, threatening to encroach her,
so she could develop out. LinkedIn has a very similar
story in the early stage, which we went through a little bit before, but we’re gonna go through
the village stage of it now. And then the question
is, usually for scale it’s a competitive
thing, so it’s relative, it’s can you scale faster,
can you scale quickly with it? And so, these are the kinds of things that get to the question of, should I, in fact, attach
a rocket engine and go? So, you’ve got a rough
sense, and by the way, each of these things is
actually more of a judgment. It doesn’t mean you have to say, okay, I have 100% checked
off each of these, because you might go,
look, I’m 50% confident on the first one, I’m 60% confident on the organization setup
in order to be able to go. I haven’t run the process so I don’t know, but you may still, because
of competitive circumstances or because of a market opportunity or because you think,
look, once people begin to realize an opportunity that’s of size, everyone starts pivoting to it. So, part of the very first
lecture’s concept was figure out something that’s kind
of contrarian and right. Usually you can do the contrarian
thing in the family stage because a lot of other people
are staying away from it because it’s not obvious to them. There’s some reason, for
example, in the LinkedIn case, like, how do you get to critical mass? How do you get millions
of people into a service, when the whole value is
the number of people there? So the first person of value,
second person of value, so they’re like, that’s not
even an area worth looking at. Once you begin to get
to the village stage, you begin to proven out revenue, you begin to proven out some of your thesis. Now everyone else can see
more of the same thing and it becomes less
contrarian ’cause they go, oh, I’m beginning to look at this, and beginning to look at this, I’m going, oh, that could be valuable. And once you see that as clearly they’re also seeing it
in some version too. And so even though you may
not have the competitor standing right here, rushing,
jostling you for the prize, you actually end up on a clock anyway. And so that’s part of the
reason why you tend to go to, okay, what’s the right scale to go to? Can be driven by competition,
but it can also be driven by the fact that now everyone’s seeing this value proposition
and you need to do it. Other reasons you might do it are you need to get to critical mass in order to make your particular
value proposition work. There’s a stack of things that lead to the decision to scale but
this is kind of the checklist for what you would be doing. – Can you say a little
bit about how LinkedIn made that choice in 2008? – So, let’s see, ’cause yes, ’cause we’re only doing
the 2008 post story today. So roughly speaking, remember, we launched the site May 5th, 2003. So this is five years in. Part of what we said at
the initial launch of it is that we said, look, let’s keep, very much actually like what
Miriam was talking about, let’s keep things trim, all right? Let’s spend only enough
cash, let’s be very selective about hiring, let’s raise money, when we do a financing
process, let’s leave at least nine months ramp
to doing a financing process because we want to
essentially figure that out. Then, basically here, we went, all right, we’ve identified the fact that now a bunch of markers say,
okay, we understand some of the core value
propositions at LinkedIn. So, for example, before 2008,
it was still pretty difficult to articulate ourselves to
journalists about what we were and so, for example, in 2003, nobody here probably
knows what Friendster is. So, there was a site called Friendster, you can just replace Facebook. The only way that I could
get in the press in 2003 was we’re Facebook, but for business. And they’re like, okay, well, we understand what Facebook
is so we’ll allow something. And Friendster was the term at the time. MySpace, maybe, probably that’s gone too. And so we got here and your
average user began going, oh, actually in fact,
I know I could use this for looking for a job,
I know I could use this to find other people,
I know that I’m staying up to date with people, I
know that having a public professional profile,
could be valuable to me in terms of someone finding me. And that was on the kind
of member value side and then on the business side,
we had begun to prove out kind of what were the
things that corporations, in particular, wanted
to spend real money on. What were the features that we had to have in the enterprise
software, ’cause we started as a consumer company and
we’re a little bit hybrid now, primarily consumer,
but enterprise as well, and so you have to figure out
a whole bunch of unique things in enterprise side like, okay, how does the account management work? How does a team function of it? How does a corporation engage with it? And all of that stuff we’d
begun to get enough of since we had product-market fit there too. And that was part of
the decision of, okay, now it’s actually time to go fast to essentially move into
the market opportunity. We want to get the consumer growth up and we also want to get
the revenue growth up and that was essentially the backdrop. That’s what you’re gesturing at, right? – Yeah, I think. Yeah, so, I mean, I
will pick up from there because I really, I distinctly remember that we had conversations in 2007, in 2008 with Dan Nye, who was CEO at the time. And we, actually, one
of the biggest topics of conversation is like,
when do we actually, we didn’t put it in these terms,
but when do we blitzscale? Because he was being very conservative about the way that we
actually went forward. So, there were four things
on the previous slide: articulate the core business,
identify critical steps, hypergrow an organization
and then finance. All these things applied
in LinkedIn’s case. So what we wanted to do, is we wanted to sort of talk through a little bit about the changes that
occurred at LinkedIn as an example of watching these
four things in action, okay? So, first one, talking about
articulating the business case. So, as Reid said, we had begun to see the growth, so we basically had proved
out that growth hypothesis. We had begun to see the value,
especially among recruiters, so we’d seen the value hypothesis and it was already out there. So, basically we were at
a place, we were like, okay, so now we need to create
a self-sustaining business, which is actually going
to not only grow the value for both those things
and continue our growth, to generate revenue and then
build competitive advantages for us over time, and then finally grow strategic assets for the future. So I’m gonna take you
through the presentation that Jeff Weiner gave. In 2010, at our All-Hands. So, 2010, we’re right about
halfway through this stage, the village stage. And basically, he presented in February the product plan or business plan overview onto the whole company. So, I’m gonna try to give
a view of a couple things I want to point out about
this as I go through it. Jeff had joined the company
at the beginning of 2009. Reid’s gonna talk a little bit about how we picked a new CEO
and why it’s so important when you’re making this transition to have the right CEO in place. Jeff, one of the things
that he wanted to bring to the table was the foundation
for growing very quickly. That meant that he needed to
communicate very differently with the company and the
people in the company. So what you’re gonna see here is something which is a far better,
sharper articulation of what this business looks
like than we ever had to do in the OS1 or OS2 stages. You never need to be this
clear when you’ve only got 25 people in a room working together, but you never need to say
anything to that group of people. But when you have, and
when the speech was given there were 480 people in the company. When you have 480 people,
you need to talk differently. This is actually the very
beginning of a set of all hands which are still going on
every two weeks at LinkedIn even as we approach 10,000 employees. So, I’m not gonna do a
Jeff Weiner imitation. I almost brought the video,
but it was too difficult to sort of bring it in. For him, articulation,
this begins with a mission, in other words, a
measurable way of saying, this is what the company’s
actually trying to achieve. This is still the same mission
that we work with today, six years on. So, connect the world’s professionals and make them more
productive and successful. It’s important to have this
because this is actually the touchstone, which becomes
part of daily decision making among the people within the company. When you get 400, 1,000,
2,000, 10,000 people in a company, the key bid is,
do we have a common language which we can actually use to
be able to make decisions, to make sure that we’re on the same page? If you actually achieve that,
you’re gonna achieve this, which is the vision for the company, which is creating economic,
well, it’s really interesting, so one tiny note, Jeff actually realized these words were wrong
during the presentation and he stopped the presentation,
he unloaded the PowerPoint, he changed the words in front of everybody because the proper term is
create economic opportunity for every professional in the world. If we’re able to achieve the mission, this is what we’re gonna
do, this is our true north as a company. Again, a really interesting
rhetorical tactic, to be able to make sure
that the whole company’s both facing the same direction. Then we talk about the strategy. – Actually, one thing, like
when you begin to think about the communication, so
you’ve begun to move past an organization of 100
people, 50 people, etc. There are now a whole,
like, if you’re a founder of a company or an executive, there’s a whole bunch of
conversations that are going on that you’re not part of. How do you have those
conversations largely resolve in the right area, where
everyone’s coherent and coordinated across it? Part of the reason you have
to articulate this stuff is because then they go, okay,
this is what we’re doing, so even though we’re not talking to that other group right now and we’re not gonna interrupt
the work they’re doing ’cause, oh, by the way, if you interrupt
so they cross communicate, then no work gets done, (laughs) right? So you go, okay I’m not gonna do that, but I understand that we’re both actually heading towards the same true north and I imagine what they’re
doing given that’s a true north this is what they’re doing
and I can crosscheck it and this is what we’re gonna do here. And that’s the reason why this kind of thing is so important. – Yeah and there’s a whole set of skills which an executive brings in,
executive who has familiarity with running a big organization, which took me forever to learn, primarily by observing Jeff,
because basically he made these words, the ones
we’ve seen on these slides and some in the future
slides, literally a part of how he managed every day with every person who
he actually spoke to. We actually repeated this entire section of the presentation at every All-Hands for a year after this started. So, connect talent with
opportunity at massive scale. This is the key strategy. So here was the understanding
about our core business for LinkedIn. The core business was gonna
be, we’ve got growth of people. We’ve got product-market
fit with the recruiters. Basically, we’re gonna
build out a large-scale and growing business for hiring, we’re basically hiring individuals. You have a large pool
of passive candidates, rather than active candidates and we’re gonna do so primarily
by driving a sales force, so enterprise sales rather
than by driving walk-up sales or basically people who are coming up and buying on the website. That was basically the structure. There were two additional pieces. One, continue to experiment and drive towards greater engagement. The greater the engagement
across the network, the happier we were all gonna be and that comes out here. And the second thing is
there were a set of secondary monetization strategies,
which went in place. But this was the basic idea. Connect talent with
opportunity at massive scale. We don’t really talk about
this any time so much now because we’ve moved past it. Our competitive advantages,
these are the moats, as Warren Buffet says. These are the moats that basically he’s asking the company to broaden. So, first of all, focus on business, rather than focusing on the personal. Continuing to grow to critical mass, remember we saw those pictures
with people holding numbers? Well, we had 60 million people by the time he actually gave this presentation. Build network density, people among it. Make sure we concentrate on data, specifically on profile completeness. Primarily strategic objectives. Continue to focus, double down
all the time on collecting, on basically having
information on professionals in order to drive our recruiting business. Second, drive engagement through
being a source of insights. And this was gonna be
our business opportunity. He drew through this
bullseye on every whiteboard in the entire building. I know it’s a little hard to read. These are actually
screengrabs from a video. So, hiring solutions was
the center of the bullseye. So, we had a very specific focus. We had lots of businesses
we could get into, but we had a specific focus
we had to get right first and that was gonna be hiring solutions. Then surrounded by a set of
marketing solutions products, which is the outside three rings. And that all sort of boiled down to a set of operating priorities, which sort of starting from the top is build a word-class team,
scale software and hardware, focus on products, monetize, and then basically expand globally. So, what Jeff was essentially doing here is he was articulating
what the overall vision that our core business was gonna be and establishing a layer of culture and a way that we’re gonna work
together from that point on. Yeah, please. – [Voiceover] When you
guys saw this presentation, what are things that were
surprising in the sense that you guys did not think about it when you were starting
(mumbles), but Jeff did. – So, easily, so the question was, what are the things which
were most surprising when Jeff came in and did this. So, one thing which was
surprising was the clarity and sharpness with which he
expressed all this stuff. We had attempted to do that before, but it was usually experimental. Jeff made it absolutely rock solid to the point where this
literally is the communication used internally, used externally, and pretty much, if you talk to reporters, they’ve all heard this
because it’s such a clear and easily repeatable message. Second was focused on culture because this is effectively a
mechanism for driving culture because we’re going to expand. Reid’s gonna touch on this a little bit in terms of creating a culture around, LinkedIn’s culture, it came up when we were comparing it with Google’s. Google’s culture may be
about highly intellectual, college or university. LinkedIn’s culture is a
change-the-world culture, which is about creating
economic opportunity. And so, in some ways, aligning behind this as well as our cultures and
values, as we’ll come to later, is exactly the way that
he wanted to be able to create the culture
which would be transmitted from generation to
generation of employees. – The key thing, and this
is the difference of scale, the way that I had been
articulating things before, which is, by the way, an
accurate, strategic descriptor of some of the stuff that LinkedIn does, is network is a platform. Your identity, the people you
connect to, real identity, is a platform for a number
of different applications that help individuals and
organizations navigate the world, by how they find each other,
how they match with each other, how they communicate with
each other, you know, how they essentially form connections too. That was the, essentially,
previous structure by which most strategy was articulated. One of Jeff’s insights is he said, look, that works for you, development team, and you geeks very well,
for articulating that. I’m gonna try to build a
company of thousands of people. I need to have something that actually, yes, we’re still doing
that, we’re still treating a network as a platform and
an application structure as kind of what we’re doing, but this is a way to rationalize
all of the different work and have people understand it. And that’s part of the
reason it was a shift. And part of the whole thesis
for understanding this is that actually, in fact,
there are ways to fail at being a 450-person company that are I just don’t understand how to scale. And so when you look
at the current giants, they’re not just there ’cause,
oh, I got the right app and had the right product-market fit and now I just hang onto
it and kind of just do what any smart person does. There’s a bunch of art in how
you build these companies. And one of the things that
both Allen and I learned from this was to say,
okay, what’s a culture that defines parallel action
across a large number of people and what’s a way of articulating the plan that most people know how to
coordinate with each other? – Go ahead. – [Voiceover] Which of
these things would be inappropriate in OS2? Why would this slide be premature
if we were still in OS2? – So which, if any of these
things would be inappropriate in OS1 or two? – Let’s look through them. So, here world-class team means something slightly different. Like actually, and part of what
he is actually talking about is a shift in culture from
what we were doing before, which is we’d hire very smart people, but part of what he’s trying to do is say, okay, look, we’re gonna actually,
in fact, operationalize. For example, now, when we’re hiring people we onboard them through this
way and we’ve created a video that essentially integrates them. We’re actually moving from, essentially, executives who are managers of managers to executives who are managing managers who are managing managers as kind of a way that we’re doing it. And so all of what’s code
in build a world-class team is adopting those practices. He wasn’t like, oh,
previously you were gonna go to the local Safeway and
just recruit anybody in. That was your previous organization, that’s not this is a contrast to, this is we have to focus
on those scale mechanisms in terms of what we’re doing, in terms of how the
operational process works. And know people to do that. – And it also gets down
to individual employees who you hire. There’s a set of employees
who are extremely good in very small environments and startups and those people tend to be generalists, they tend to be people who
love the startup environment. They tend to be people
who want to experiment and work all night, take huge risk. And then there’s a set of
people who basically work much better in an environment which looks like a big company. So part of building a world-class team means hiring the right
set of people for scale, whether you’re an executive or not. I don’t know if you
wanted to say something about generalists, specialists
in the transition here. – Well, and this again
wasn’t on Jeff’s mind, nor perhaps this theme was not on my mind until I started constructing what this kind of playbook looks like and why this is actually,
and to some degree, the visible secret of Silicon Valley, but you hire generalists
the whole way along this kind of scale up, but
what you’re doing with them and why you hire the generalists and what the generalists does is different because you’re getting
a lot more specialists that do different things,
including, for example, specialists who are managers or specialists who are executives. And so, here, part of the thing is a lot of what you’re key
pivotal leadership roles may be occupied by generalists, either they’re gonna
have to be specialists and then start doing
that role as a specialist or they’re gonna have
to move to other roles. Because part of what
generalists allow you to do is be flexible, attack a new
problem, figure something out. We had a number of them
organization, we still do, but when you’re saying, okay,
what we’re really actually trying to do is take this
organization and now grow it by 300% and have it
operationalized on a certain set of metrics and dashboards
all while improving its operational efficiency
while doing that. That tends to be people who
are very expert at that game versus the people who
are kind of going, well, maybe we should do it this way
or maybe we should try that or maybe we should try this and, you know, that kind of thing, which
tends to be more generalist. And so that’s the (mumbles). Although that wasn’t, that’s
a truth about scaling, that’s not necessarily what
Jeff was saying on this bullet. – On the second one, I mean, for scaling hardware and software, one of the things in the OS1 and OS2 phase is that generally speaking,
what you’re optimizing for with your technology is
you’re optimizing for agility. So, basically, you want to be
able to create an environment where you can experiment
with the nature of things, learn effectively, go try something else, just abandon previous
failures and so forth. When you actually get to
this stage, you need to shift because what you need to do is 10x or 100x the technology infrastructure
that you’ve actually got. You need to get to the
place where you can support a massive, scaling thing. So the attitude towards the software you build is quite different. – Yup. ‘Cause we were trying to essentially say, let’s figure out what are the core set that is in network is a platform,
so we were building a lot of what we called .2, .3 features. We’re trying this, we’re trying
that, that sort of thing. Here was to say, okay, look, we’re gonna focus on the things
that are currently working and we’re gonna scale
those, versus doing that in terms of a prioritization. It’s not to say we don’t
still do some of those, but this was the kind of prioritization. And that was also part
of what got to the third, sorry? – [Allen] He had a question.
– [Reid] Oh. – [Allen] Just tell him one second. Please finish your thought.
– [Reid] Oh, okay. – [Reid] The third bullet, which is, okay, what we’re looking for is
a small number of products that we’re really going to
change the scale characteristic as opposed, ’cause we
believe we already have the product-market fit for the key thing that we’re gonna scale up on because while network is platform
was the general strategy, hiring is only one application,
but we’re gonna use that to drive all of the scale. – And that’s our initial core business that we’re starting from because the strength in that business allows us to build all the others. A question?
– [Voiceover] So going back – [Voiceover] to the
generalists, specialists divide, how do you both determine if someone’s a true generalist or just a bad specialist? (Reid and Allen laugh)
And how do you know then, if the generalist can become a specialist before they fail or succeed? – So the question is how do you figure out true, high quality
generalists and how do you distinguish between
generalists and specialists? So, high quality is a
judgment art on both, ’cause sometimes, by the way, the high quality specialists
are also hard to figure out, ’cause sometimes they’re specialists in something other than you, ’cause by the way, when
you’re doing the family side, basically everyone’s generalist. So, first is, a good
generalist is somebody that tackles a new problem, has a relatively good risk triage, knows how to sort it out, is
comfortable doing something that’s fairly different than
the thing they’ve done before as they’re doing it, so, for
example, a generalist engineer who might be good at servers, ’cause okay fine I’ll try
the iOS frontend and so forth even though I haven’t really done that, I’m really a specialist
in the other thing, just as kind of an engineering
parallel or CS parallel. And then a good generalist usually has a very high analytic clock speed, an ability to learn stuff very quickly, and an ability to recognize fairly fast when they’re not on the
right track, so they correct. It’s not that there’s no foot faults, they’re gonna go, okay, right, over here. And if they don’t have
those characteristics, they don’t tend to be a good generalist. They may be a good specialist, but they don’t tend to
be a good generalist. Specialists, you then tend
to evaluate more the way that when, for example,
you’re thinking about hiring, about like, okay, what’s
the specific kind of work that you’ve done like this before? What have you learned in that? What is expert on it? How do you reproduce it? What is the way that that’s applied here, whether that’s management,
server operations, product management, sales, etc. So like, for example,
let’s take a sales example, if I’m hiring a sales generalist, I’m kind of looking
for, okay, what happens if I throw you in this circumstance? Like you have this kind of
weird client, what do you do? You know, kind of like
different circumstances. When I’m hiring a sales
specialist, I’m going, okay, what’s the way that
you define a territory, how do you manage a sales pipeline, what’s the way you know your sales pipeline’s actually working, and a set of things that come
to the I’ve done this before and I know these metrics very well. – [Voiceover] Do you
internally classify people as specialists or generalists and do you interview differently for those different roles? – I definitely interview
differently and generally speaking, yes, sometimes it’s a little murky. So it’s not 100%, but generally speaking, I’m either looking for
someone who goes, no, no, the job is doing this role,
really killing those role, yes you may grow to other things, but this specific thing,
which is actually, in fact, something that you
could’ve learned before, or no, actually, in fact, we’re
trying to figure this out, not many people know this. Like, are you capable of
solving, and by the way, sometimes you still have
to have the toolset. Like, for example, let’s
say a generalist engineer versus just a generalist. You still have a toolset that allows you to free roam within that area. There’s a… – [Voiceover] Throughout
your time at Linked, did you make any bad hiring decisions and if you did, what
did you learn by that? – Lots (laughs). – Did we make any – [Both] bad hiring decisions. – Yeah, we hired this guy Reid. (class laughs) So, you have a couple of different choices when you’re scaling, in terms of hiring. So one is you can go, we’re
gonna hire relatively easily, but then we also have to fire fast. So that’s one way you
actually do an organization. The other one is, you have
to hire much more carefully and you still always
tend to fire too slowly. You tend to be, you should fire faster, but you’re kind of more careful on that. Most of the organizations that I tend to participate in,
tend to do the second versus the first because
you’re trying to build also a community and a team and
if you’re churning people out or it kind of feels like
the ax falls pretty fast, it’s very hard to have
that kind of teamwork fall. There are high performance
organizations that do do the first pattern, though. – The only thing I would add is that one thing you have to take into account is that movement from stage to stage. So when you hire somebody
for a specific stage, you know you’re moving
on to the next stage and the next stage and
you know the requirements for what that person can do
is gonna change over time. So the person you bring in for OS1 and OS2 might not be able to make
the transition to OS3. That doesn’t stop you from hiring them. You still have to try to
hire the best person you can to get to the stage you’re in right now and then hope you can grow
them in the next time around. – And part of the whole
point of what we’re hoping you guys in the world
is learning from this is when you hire someone,
for example at a tribe, and you go, look, I don’t
know if they’re gonna get to the village level, right? You’ve started thinking
about what that question is because you’re starting to gather data even when you’re at the tribe level because part of the question
is, you may have a person you really wanna keep who
is no longer going to be the head of that function when you get to the tribe level or when
you get to the city level. They’re not gonna be there anymore and so, your only hope of keeping them is you start that conversation early. If you walk up to that person and say, look, I really know you think of yourself as the head of engineering, but really, (Allen laughs)
meet Sue, meet Bob, they’re the new head of engineering. They’re lost (laughs). They’re not gonna stay
in the organization. And so part of being able
to have the lens ahead, and by the way I will also get back to the bad hire question, but when you get to the lens ahead and you’re thinking,
I know this is coming, is this person likely
to be able to get it? It’s almost like an investment thesis. And if they’re getting it then
you’re giving them coaching and you’re trying to get them there, but then sometimes you’re doing is saying, look, I’m trying to preserve
them in the organization, I’m trying to get a path
that makes them feel like their own personal mission is being done, but they understand why I’m gonna go hire a new head of engineering and that that will be a collaborative effort. And understanding going into that means that there’s very good
people at, you know, say the tribe of going to the village that you’re then trying to preserve. You will not preserve 100% of them. Some people will go, my whole
identity is I’m the head of X and you’re taking that away from me and that’s not gonna work. And by the way, again,
when you’re thinking about this lens in advance, you go, oh, that’s where their identity really is and when I get there, I’m
either gonna have someone who’s not gonna work and I’m
gonna put them in temporarily for that ’cause it’s not gonna work, or I gotta be planning
on trading them out. And that’s part of how you’re
thinking about growing the org and that’s part of the
reason this is real work and not just kind of like, oh, well, look, we hired another 100 people. – For those of you who
are working on hiring for startups, though, founders
are a different class here because founders always
associate their success with the success of a company and not with their
success of a given role. So, basically, founders tend to be willing to make changes later
on in order to be able to continue to contribute over time. That is not necessarily true
of people you would hire at the tribe stage or beyond. – The short answer on the
quickening of the cycle on the bad hire decision is, the heuristic that I frequently use now is now that you get more data,
would you actually, in fact, from today hire them again? Now that you’ve worked
with them and so on? And if you don’t actually
really answer that question yes, start figuring out a
graceful parting of ways. And that’s the kind of
trick is like, no, no, it’s just like stocks. Like, okay, no, no, a
hold is a buy! (laughs) What I buy now, right? And that’s kind of a proxy for
all of the different things that go into is it a good
hire, is it a bad hire, is the chemistry good,
do they understand this, do they have the right risk tolerances, do they understand
sometimes you just kind of, and by the way, this is true for everyone in the organization,
including me, you know, at these stages sometimes
you just shut up and row. (laughs) Right? You go like, okay, we
need to get through this and understanding that. And there’s all kinds of
things that lead to that. There was a question also, yes. – [Voiceover] So, 60
million users at this point seems like a lot, so why did
LinkedIn not have to scale, before or why did it
not scale before that? – Well, so one of the key
things was before we knew, ’cause remember you have
to scale in capital, before we knew that we would
be able to massively grow the revenue line in consonant with it, we were always trying to
grow the overall user base and we’re using techniques for that, but until we knew that we
could actually, essentially, get a strong revenue capture,
reinvest that revenue in the general consumer
internet, and that would happen, then part of the risk you make, so, for example, one of the reasons, like I wrestled a lot
with the term blitzscale and the reason was is
that there’s some good and interesting intellectual parallels to the term blitzkrieg. People look at blitzkrieg
and they go, World War II, Nazis bad, (laughs) right? You kind of go, okay. Not ideal in terms of of the term. But actually, in fact,
there’s some real parallels and one of the parallels,
one of the key innovations that was made in blitzkrieg was that prior to blitzkrieg, all war was done through a supply chain. So, it was, you only extended your front as far as your supply chain could do it. So you had a maximum speed
you could advance your front. What the essential
innovation in blitzkrieg was was, eh, heck with the supply chain. What you can carry, go fast. What blitzkrieg ultimately meant is once you got to the halfway point to your intended battle, at that point, you cross that, you win or lose big. If you lose, you lose terribly ’cause you have no supply,
you have no backup, you run out of ammunition,
you don’t have food, you just gonna collapse
when you get there. So the halfway point, you’ve kind of made this point of I’m gambling,
or in the parallel that I frequently use for startups, I’m jumping off the cliff
and assembling an airplane on my way down. And similarly, when you’re
beginning to do the scale, you’re like, okay, now I’m gonna really crank up my burn rate. I’m gonna hire a whole bunch of people, I’m gonna really make a go with this. And if you’re wrong, it’s pretty painful. It may be death. If it’s not death, it’s a massive retrenchment loss of opportunity. And so it was like, okay, let’s make sure that actually, in fact,
we have a capital engine through the revenue that
actually can do the scale up. And that’s essentially
what we were looking for. – So last question and then we’ll move on. – [Voiceover] So, it’s interesting
that you talk about Jeff (mumbles), you never mention Sheryl. Do you have a thesis
on whether bringing in that outside leader is a critical element correlated with companies
that make this leap? Or whether the old team
can adapt on their own? – So– – You wanna repeat the question? – Oh, yes, thank you. So, the basic question is, look, we’re talking about Jeff Weiner, who we brought in at the village stage, we’re talking about Sheryl Sandberg, who helped with Facebook and also was brought
in by Eric into Google. And what’s the thesis about growing the team versus external? So, roughly speaking, I think that your true north always in these things is some combination of external and internal. And if you try to apply too much of a rule one way or the other, you break. So, if you say it’s all internal, you tend to only drink your own Kool-Aid, unless the people had a
lot of experience themself with directly scaling before, which, by the way, is a
relatively rare experience because the number that go
from here all the way up is actually, in fact, relatively few. Plus, you may not keep modernizing, you may not be figuring out what does the current market look like, what does the current
talent market look like, what does the current
go-to-market strategy look like? And that new blood is actually
very helpful for that. So, people who go all internal and no external have a problem. The people who go all external
tend to lose a lot of people who are like, well,
what we’ve been thinking intensive at this problem,
you know, 100 hours a week, we’re emotionally committed to it, we’ll basically roll
around in coals for this in order to make this work. And so if you lose them,
and so the whole art is how do you figure out where it is between those two pillars, right? And some of it comes down to founders in the key team recognizing
what their key strengths and weaknesses are and what they can do and what they need to have. And part of the reason why
you have external folks like board members and
investors is to help be that kind of dialog, critique. Saying, look, maybe you
should think about this, this is why and so forth. And so for example, when
you’re founding a company, the question you don’t want to ask is, oh, am I doing all right? You wanna say, can I be doing better? Am I hitting it? What do I need to be doing better? Because you wanna have an
accurate judgment of this because if you don’t have an
accurate judgment of this, you’re likely to fail. You’d rather hire somebody and succeed. And actually, in fact,
that was one of the things that was both super impressive. And watching from, you
know, ’cause Peter Thiel, Mark Pincus and I all kind of invested in the very first round of Facebook, watching Zuckerberg grow because he was trading out execs as he was going. He was going, oh, that’s good. Oh, that’s better, okay, let’s type here. Like, for example, you
saw him kind of going, oh, this whole organization’s fine. Then he hired Sheryl,
saw how Sheryl’s running an organization, and then turned around and improved his own organization. So, it can be a catalyst between them, but it usually is pretty important to have both function and how it works. And then there’s, just on the specific kind of founder, CEO
question, there’s an essay that’s on LinkedIn and also
my blog, reidhoffman.org for that specific question. – And as a transition, the
only thing I would add here is that the reason we want through like, here’s the core business,
we’re trying to articulate it is because the next section is
about what is the actual plan that we’re actually gonna
follow in order to do it. And the key part for all of this stuff is you gotta have the
people who are gonna be able to execute that plan. And they’re gonna
execute it as executives. Basically, that’s the most
important consideration on a per-person basis. So, I’m gonna move forward on this– – Just very quickly, just a second. We were experimenting with
a lot of different things ’cause remember we’re staying agile, so this was focused on a few and that was part of the scale up and that was why it was different
to that earlier question. And then this was, we
are massively focusing on going international and
most kind of blitzscale stuff is you go global. There’s very little blitzscaling stuff that isn’t going global. The primary exceptions
to that are within China. – Yeah, and neither of
these played a big role in the first two stages for LinkedIn. So, I’m gonna go to the next section, which is about critical components. So, if you’ve got a plan,
which is basically about, like we had for Jeff,
that Jeff went through, what do you actually
need in order to do it? And almost always there’s a
set of critical components, which you actually have to focus on. You don’t have to focus on everything. So as always, you’re growing. We’ve only got 480 people at
the time we’re doing this. You can’t do everything perfectly, so the question is what are like two or three things you have to get right? And for us, those two or
three things were product, go-to-market and engineering. So, I’m gonna go through
these really briefly. One thing I wanna note
about this is that each of these three plans was presented by a brand-new executive at LinkedIn. So, all of those people were people who were basically hired
in at the OS3 level. So the first one is our product plan. I’m not gonna go through the
details on this, obviously. We don’t have time, but
this kind of organization, where we see technical
foundation, viral growth, professional identity ecosystem
insights and monetization. This kind of organization was something we had never seen before at LinkedIn. We’d never taken this analytic approach to doing anything. And, sorry, you were gonna say something? – I was literally gonna add
to that earlier question. We knew a bunch of this stuff, but we’d never construct
a document like it ’cause we weren’t trying
to figure out how to get 500 people and growing to all
be executing on the same plan. – And also, this is the very first time in the company’s history
where we had to execute on multiple threads simultaneously. When you’re a smaller
company, generally speaking, you’re gonna do one thing really well and then everything else you
sort of do as you need to do. Now you actually need
to be able to execute on three or four things
at scale all at once, which is why you need to be
able to sum it up like this. This broke down into a
set of product features. We separate things into
core, strategic, and venture. I’m not gonna go and get
into the details here but the idea was we want
to identify the things which were part of solving
our core business problem. All this other stuff was
stuff which we thought was building for the future. This stuff was about
building for the present, about making our core business operate. The product plan basically
focused on member growth, on professional identity, on search, and obviously an important component of making recruiters successful, on knowledge sharing,
which is one of those sort of future-facing things. And then our three businesses, so hiring solutions was the key one, that was the center of our bullseye. Marketing solutions and
our monetization platform, which was basically payments and a bunch of other things that we wouldn’t have to have thought about before. So, that’s on the product side. In terms of the go-to-market, we’re just gonna talk about sales. You can see the way we’re broken up. The blue bars here are
basically the places where our field sales team was responsible for driving revenue. And the others were
non-field sales related, so– – One other thing to think
about is I think about kind of how to understand this growing. You’re beginning to run
organizations like this. These are all documents that were taken from how we would communicate internally, how we’d talk to each other, how we’d formulate the same plan. You’re beginning to actually
crystallize the business within graphs, dashboards, other kinds of things as a way of doing it. And that’s something
that’s not as critical when you’re at the OS2 and tribal level. – And it’s important
for a couple of reasons. So, first of all, it’s
important because it allows you to project what the business
is actually gonna look like, which is something you actually need when you get to this place, ’cause you need to know you’re tracking toward a particular target. You need to be able do the
financing appropriately and a bunch of other things. But the second thing is it becomes part of the way the company is
managed on a regular basis. This was the very first
time, right about this time, when we started producing
daily dashboards, which basically, and
Jeff still looks at them every morning at 5:30, where basically he would run the company
based on a set of dashboards, which he would receive in email and literally look on his iPhone in bed the minute he woke up to sort of find out where the company actually was. There was a brand new discipline, which had not existed
in the system before. – I certainly wasn’t doing that. – I…
(Reid laughs) So, again, not a gigantic
amount of detail, but the key piece was breaking things up, how do we understand our global, how do we understand our markets, how do we understand our global approach, how do we understand our
geographical approach? Product mix, how do we build a field force to basically go out and sell effectively? What about brand leadership? A whole bunch of questions
which are the kinds of questions you would expect from a
seasoned sales leader, but not from a generalist
who was basically trying to make it happen. On our way to a billion dollars. And finally, engineering. So engineering was the third major area. We actually brought in a
guy called David Henke, you’ll see a photograph of
him later if we get there. Basically, building out our technology. His goal was 10x. LinkedIn has a bunch
of really hard problems and we’ve got a bunch
of productivity issues that we need to deal with. The question for us is basically how do we attack these
things in an organized way and get to the place
where we have 10x scale on top of where we are right now? – For example, on all
the productivity stuff, part of what happens when you get in here is you begin to actually focus
on developer productivity, you’re building your own tool stack. You’re leveraging other things as well, but you’re beginning to figure out how does that cycle begin to work. And this is usually where you’re doing it, whereas before you’re
just building the thing that demonstrates, do I have
a shot at product-market fit? – Yup, so the first thing
was like building up time and then we’d have this
selection of things, you can see tools in productivity,
distributed computing, security, disaster recovery,
a whole bunch of things that you never think
about in the early days, but when you actually have
a business built on top– – Or you think about thinly. – Or you think about very
thinly at the very beginning, which you didn’t think about in depth, but now you actually have to. This was actually the
largest transformation which had to occur at LinkedIn because basically we had to
take, not only are we growing very, very quickly in this
space, in terms of engineers, but also we needed to change
basically the entire stack on which they were working,
simultaneously with continuing to develop product and
scaling simultaneously. So it was three things
they had to do in parallel. – [Reid] And all scale companies do that, that’s not just LinkedIn. – Yes. This was a quote from our
original head of engineering, Jean Luc, who had moved
over to a new role, which we can cover later. Go ahead. – [Voiceover] About how many
engineers were at this point? – I don’t know, I had the number
in here, the exact number. At this point it’s about, just under 200. Sorry, that was how many
engineers were there at the company. – [Voiceover] So in the process of growing the new team,
transitioning the stack and building the product, in the process of going to market more aggressively to scale, these two process going on simultaneously, but independently Did you make sure you
weren’t selling something (mumbles) or is that not a problem? – Well– – [Voiceover] Better
to be selling something ready than (mumbles)? – So with all these things
going on simultaneously, how do you make sure you’re
not selling something that you don’t actually have ready? The short answer to the question is the whole reason that
we brought in a process where we could concentrate
simply on three things that have tight integration
between those three executives, so just to make sure we
didn’t lose coherence between those three things. We had a very specific pecking order between the three of them. Product said is what we’re building, technology said this is what
we’re actually able to do, and sales sold only the things
that were actually produced. But it was a very clear understanding of what the cascade
actually was through them, which actually made that system function. Okay. So I’m gonna turn it over to
you to talk about organization. – Yup. You know, when you begin
to say okay, village, OS3 has been where you really begin to identify the scale challenges. This is one of the central
places where you go, okay, do you have the right CEO? Are you the right CEO? Do you have the right CEO? Because you’re not just thinking now, you’re thinking what happens
when it’s 1,000 people, 2,000 people, 3,000 people. Because the mistake people frequently make in organization stuff is
they say, okay, let’s wait until it’s all broken, then try to fix it. That’s much harder.
(Allen giggles) Part of what you’re
trying to do is anticipate where it’s going and figure it out and then make the adjustments
while you’re going, whether or not this individual
who works well at tribe, are they gonna work out in village, what’s a role that’s
gonna work out in village? CEO is obviously central to that because a CEO is gonna direct, this is how the whole
thing works at scale. And it’s one of the
places where most often you either have to come up
extremely steep learning curve ’cause you really, really want to and this is the thing
you wanna do as a founder or it’s the beginning of the time saying, okay, do I do it by getting a CEO? Do I do it by getting a COO? You know, do I learn some of it? What are those kinds of things? Those are kind of the key things. And then you begin to
look at kind of what are essentially, the key executives. Now, part of when you’re
looking at this stuff is you don’t try to
solve all the problems. You don’t go, I have a perfect org chart, I got six boxes, I
gotta fix all the boxes. Every single time you have a list, whether it’s problems or
other kinds of things, you should force rank them and you go, okay, what are the ones that
I absolutely have to do? Let me get those done and let me try to do the other ones
too, but let me do that. And then what happens if I
only get to two of the six? What am I gonna plan on doing? Now, part of the role of generalist and people who have been
around the organization a certain amount of time, is that sometimes they can fill in on some of that stuff and that was, like, for example, one of the
things that I had been doing, you know, for, what was it? Two years, was I head of product? Something like that?
– [Allen] Mm hmm. – [Reid] Anyway, something
along those lines. And then you kind of look at, okay, are we capable as an
organization of doing it? Here is actually frequently where you may have hired your
first HR person at tribe, may or may not, but now
you’ve got an HR function you’ve actually got to get to grow and it has to work well
and has to be partner with all the different
functions in a good way. You have to actually, in fact,
now start thinking about, okay, it’s not just we all know it, but we actually have to have
some documentation of it. We have to have some understanding of how, so people aren’t surprised
by the people’s work and it doesn’t trip them up. And then you have to be thinking
about what does this mean for the early organization
in terms your doing? And how do you essentially use dashboards and metrics as a way of doing it? And so, one of the things that I had done is I had hired, ’cause I knew pretty early that I’m good at product and
business strategy things, but didn’t actually wanna grow a large organization and be the CEO. I’d hired a guy named Dan Nye. Dan did a spectacular job. He basically said, okay, got there at 65, let’s start actually, in
fact, putting more structure around this versus just trying to solve, only the product-market
fit problem had grown it. But then one of the things, when Dan and I were
talking about, we realized one of the things that was
really core to the strategy of a scaling, large organization was to say, you have to have an affinity with the exact problem. The CEO has to essentially embody this problem’s product-market fit. And so, Dan was kind of a
classic excellence leader, a class enterprise person and so forth, yet we’re a consumer site first. And so when we were
talking about that we said, okay, what we’ll do is we will
actually have me step back in the CEO, it’s like look, but I’ve got all these organizational
kind of things where, organization process does
not work spectacularly well under my direction because
I’m always getting attracted to whatever the new
most important fire is. Allen has been burned on
this a number of times. (laughs) Right? And so we’ll bring in Jeff to help me. That then very quickly moved to, oh, Jeff should be the right CEO and then we put Jeff in as CEO. And this is kind of a way of understanding kind of how quickly you had to be working on a number of different
problems to keep a organization with high cultural morale and coherent in terms of doing this. And this one in particular
shows these are all the people who essentially, all the red was in all these different functions for recruiting were the new people. – And just to be clear,
this was joined in 2010. This presentation is from May of 2010. – Yeah. – So this is the first
half of the year hires within the recruiting team. – Yup, and so, that’s
because the question is as you’re scaling this team, it’s actually nontrivial to be
getting the right people in, keeping the right culture,
keeping the right communications, and there’s a whole bunch
of work that goes into that. And so these are kind
of some of the topics that we were essentially, this is, again, pulled from internal slides that we were like, okay, here’s
the thing that we need to do in order to be able to work well together. And this is kind of what
the hiring ramp looked like. And I guess that’s when we really started our real big intern program too, as I now think about it.
– [Allen] First year of it. – [Reid] Yup. ‘Cause we were trying to figure out how to continue to scale
and the intern program was a way of doing that. So, do you wanna do this or?
– [Allen] Yeah. – [Allen] So, briefly,
you can actually read this on LinkedIn in completion. You go to Jeff’s profile,
you can find it from there. Basically, the idea was
that Jeff knew from the very beginning that we were
gonna grow really effectively. One of the things he had
to make do is make sure that the values and culture of the company were communicated via the hiring process. So, then basically we as a company, when we were hiring new
people, in a distributed way, every person who was doing hiring was using the same bar
for culture and match and so forth over time,
otherwise we’d end up with a cultureless company. So, what he did was he articulated
the vision, the strategy, the mission and so forth. Go next. But he also articulated this
set of culture and values and you can see here
they’re stated in the sense of here’s how you use
this to test somebody during the hiring process
because he need to make sure that was part of the way
we were thinking about each new person who came in the door. – Yup, and frequently, we
actually have a number of places where this is on the wall so
that people can refer to it when they’re talking amongst each other about how to make a decision on a values point or something else. – Yeah, and it’s the rooms
where we actually hire people. The key thing about this though, in terms of making this work, and Jeff pointed out in his post, is you can’t just talk about it, you can’t just put it on the wall, you’ve gotta make it part of the way you manage on a daily basis. And to their credit, Jeff
does it on a regular basis, but the person in the company
who actually does the culture and value stuff best is Mike Gamson, who’s the head of sales. And basically he runs
our sales organization very differently than any
other sales organization. He runs it based on this set of values. Okay, so I wanna talk a
little bit about process. Some notes, this is from a tech All-Hands. Basically, people recruiting
was the number one thing that the tech team was
actually worried about. So community in that
culture was essential. There’s a ton of stuff here
about what would be great about coming and working at a place and actually some
reasonable advice in here about what attracts tech
talent to a particular place. – ‘Cause by the way, part of
the thing that this underlies, which is one of the reasons we decided to even put a kind of an
eye-sore slide up that, don’t bother trying to read
out all of it and all the rest. – Yeah, it’s on the video. – But the thing to understand was, so when you’re a small
company, you’re a tribe, you’re something else, actually,
in fact, part of your pitch to new talent coming in is
be part of the early set, help define this, help find
the product-market fit, be one of the first engineers working on iOS or Android or something else When you’re scaling up, that
pitch doesn’t exist anymore. So you’re changing your
pitch about saying, how do you say, who is
a high quality talent that you wanna have join. You’re evolving it. It may not change 100%,
but you’re changing it. And this was part of
the, okay, how do we need to now talk about what is the mission that high quality folks,
technical talent, engineers, would wanna come and say work at LinkedIn. And that was part of
the work that was done as part of enabling the scaling. – Yup. So, I didn’t realize this had build in it. Of course I have to go
through it all again. This is a set of things which
was a part of a discussion the team was having ’cause we moved into the world of process. Again, you don’t need to read all of it, but basically the team, which is now a little over 200 engineers at the time this slide was written, were basically in a
position where we couldn’t effectively work together anymore. We had too little structure,
too little understanding of the way the pieces came together and we had a build system
which looked like this. Or sorry, release system
which looked like this. And believe me it was just as painful as this diagram is to read. We spent years chasing
after the right kind of release system before
we got to the place where we usually have DevOps today, but this reality is where we lived. Getting these processes
right is absolutely essential to letting 300 or 400 or
500 people work together. Okay, so… Sorry. Oh, that’s David Henke.
– [Reid] Yes. – [Allen] And he’s reminding
us that when we have a spike like that, that means
the site’s throttling, which was happening back in 2009, and that that can’t happen anymore. He looks angry, right? – [Reid] Yup. – [Allen] All right. – All right, so, the other
part of going in and scaling is to make sure that
you have enough capital. And so one of the things
we said is all right, we’ve got money in the bank, we’ve got millions of dollars in the bank, but we’re about to throw in the gas. We think we have a good
scale of product-market fit. You never are at 100%. You’re always at some percent, you know. We probably thought we were at 85%, but, you know, what if
you discovered it was 60 or takes an extra year or so? So we said, okay, what we
should do is we should raise enough capital to smooth out a challenge. By the way, doing that before
you’re in the challenge is actually much less painful, again, than when you’re in the challenge. And so I went out and kind
of raised this series D, which is the last one. Now, many of you may be familiar with a LinkedIn series B
deck, which I published. And you should, if you’re
not, you should be. But this is essentially,
like, there was a deck, which I will go through
a couple of the slides, but part of what happens when you’re actually getting to scale, you’re able to articulate
your pitch very succinctly. And this was like, this
is what the company is, this is what we’re doing,
this is why we do it. So, when you go down here
you say, for example, why do you have an email address? Well, that’s because,
look, our viral engine actually, in fact, if people
choose to invite people we can actually still be
growing and inviting people. It’s not actually, you
know, kind of tapped out. And we have a number of
people, who are all connected, interacting with each other,
so that’s likely to happen. Here’s a summary of what
our revenue growth has been. And so you go like, okay,
where does that lead to? And you kind of build a model
and you talk about that. You say, look, we’ve got
very smart people behind us and then you go, okay,
here is the key thing about why this an
interesting business model. We essentially have no cost
to customer acquisition, we’re growing high, we have a high margin because it’s all digital goods. You go it’s highly scalable because we can just replicate
it to as many people as possibly can happen. We have network effects
on what we’re doing and there’s huge markets
in each of these things ’cause this gets back
to the network and apps, like recruiting is one
app, but so is kind of what is business media, what
is services, what is sales. There’s a bunch of things,
obviously a bunch of this stuff, you know, who knows if we’ll ever get to. And then the other part of pitching is when people are like,
look, I haven’t really studied your business, but I wanna
know why to be excited. It’s the analogy and this was covered in the series B, as you say, look, here is some value companies and this is how we’re like them. Doesn’t mean we’re gonna be exactly that, but this is a way of going, oh, right, if you’re right about
one of those analogies, I wanna be an investor,
is essentially the pitch. And so this gets back to, ’cause remember, we had shifted and part of the
reason we included this slide is we said we’d shifted from
how we were articulating it from the internal plan we were managing, but this was still the
network is the platform was the key way that we’re
driving the strategy. And part of the thing
that we’re gonna get to is say, okay, how do we essentially, look, we’ll already in this thread, this is the capital that
we’re trying to raise in order to really
scale our business model and get this compounding. This is kind of the key path to doing it. Then you kind of get to, all right, so, is it a bunch of kind of random people, your numbers look big,
but they’re not that good. So frequently when your pitch, ’cause you’re thinking
about why would they want to and what would their objections be. Part of their objections
is do you have good people and it’s like, yes, look, here’s the thing you can send your analyst to go check about what’s going on and this is, what? 2008? Yeah.
– [Allen] 2008, yeah. – [Reid] Right. By the way, part of the
reason to include this is we were saying, look, we can get more into the media business
than we currently are. This was 2008, this is well before
influencers or anything else. Most of these strategy
plans tend to be long laid. They don’t necessarily,
each tactic is not thought, like influencers was
something we thought about relatively longer, but we
said, we have a network as a platform, we have a
personal identity as a platform, what kinds of things we can do. We can help people figure out what kinds of business information
is most relevant for them. We included this slide mostly to show that as you’re beginning to scale, you’re doing this kind of analysis. As opposed to I have a
concept, you’re saying, here is what mean revenue per member looks like in each of these different circumstances, here’s what we think we
can possibly capture, and then here is current
revenue projections and here is if we get certain things right and we talk about those
things being right. Actually, it’s funny,
we haven’t done this. We should go back and crosscheck how much of the upside, yeah, we should do that. I haven’t done that. I mean, this is a long time ago. So, that’s essentially,
kind of the LinkedIn going through the village and prepping for the really major scale and you have to solve the capital thing, you have to solve the
organization and talent thing, you have to solve the communications, you have to figure out what’s the way that as you’re growing
all this organization, growing your fit-to-market,
that actually, in fact, you stay at a strong
fighting organization. And that’s part of the thing
that is actually, in fact, part of the core challenge. – And you have to have clarity in terms of what you’re
actually trying to get done because it’s very easy
here, as in any other stage of a company, to get too much
focus on too many things. So, getting down to a single thing and making sure that it’s right
and doing what’s necessary for that is a key part
to actually navigating this section of a company’s
history successfully. – [Reid] Questions? – Oh, please, you go ahead. – No, it doesn’t matter. Go ahead. – [Voiceover] So how many companies or what percentage, roughly, fail at this stage and actually is failing at this stage fatal or does it just mean
you need to try it again at a slower rate of changing
some of your assumptions? – What percentage of
companies fail at this stage and can they go back and do it again or? – Well, so what happens,
it depends a little bit on what your competitive
landscape looks like for a lot of these companies, so there are a bunch of companies
that fail at this stage. Sometimes it’s product-market fit, sometimes it’s technology. Friendster, I think, fundamentally
failed at this stage. – [Allen] Yeah.
– [Reid] Right? – [Reid] As an example. – Technology in their case. – Yes, it was technology in their case, plus kind of an inability to
pull the organization together, there was a lot of infighting as they grew.
– [Allen] Yeah, very true. – [Reid] There’s actually a ton. I don’t know what the exact percentage is because we all tend to
look at the success biases. We look at the ones like this
and say, oh those ones work. (Allen chuckles)
But, you know, like part of like, you
think, Friendster’s gone, MySpace is gone, (laughs) right? There is– – [Allen] Orkut reached this scale. – Orkut, yes, Orkut was
definitely, yes I remember Orkut. Anyway, so there’s a ton. And right now we’re only
within the social space, which is a place we’re watching carefully. You tend to look at the
same thing with any commerce and a bunch of other things. So, the short answer is, there’s a lot of different failure things,
some of them is inability to manage the scale stuff. Sometimes it’s other things too, but there’s a lot of
failures at this point. – [Voiceover] How did you
approach launching a new location? Like you’re going international,
you’re starting to focus. And the viral loop you
had is very different in each location you’re going, so how do you redesign that viral loop, and how’d you make it work in the end? – Do you wanna? – The short answer is we
didn’t redesign the viral loop. One of the things which is
a bias for early companies is that globalization happens. It just might happen.
– [Reid] Especially – [Reid] internet companies. – [Allen] And especially for– – Little bit different
for enterprise, but yes. – For sure. So basically on day one for LinkedIn, we’re back in OS1, we were half international and half US and we never dropped below that ratio. So globalization just sort of occurred. That was before we translated the site or any of that kind of stuff. So, the viral mechanism
all remained the same. Now, it doesn’t function, as
you say, the same everywhere and the places where it doesn’t
function in the same way are places where our growth
has been extremely slow. – And it’s more in the
last couple of years we’ve started actually, in fact, doing the specific changes for it ’cause part of it also
is we were still getting such good growth from email,
it was like, okay. (laughs) – Yeah. – And it wasn’t the absolute top problem. – [Voiceover] So,
(mumbles) of fundraising, (mumbles) of that fact that
you’re trying to blitzscale to out-compete some of your competitors. How do you not fall into the temptation of doing what a lot of, especially in the current climate, of just raising so soon
and raising so much, just to be able to
outspend your competitors. – Can you repeat it a little bit? – Yes, so the question is when you get to a spot,
what’s to slow you down from just raising a ton of money and using a bunch of
capital to out-compete. Well, I think you have to think of a couple of different things. So, by default, you actually
raise more capital than less. So you say, I can’t get
to a coherent answer, I just raise more capital. Because the more capital
gives you more optionality. There are problems that come with it. Sometimes exit prices,
sometimes a lack of discipline in what you’re doing. Sometimes, you know, kind of what are the expectations
of the new investor? All of those things can actually
complicate your execution. So there are downsides to it, but by default you say,
actually, a bigger war chest, intelligently managed, is
better than a smaller war chest. Now, one of the prices
is obviously dilution. So depending a little
bit on what the price and everything else happening. But part of the reason,
the last couple years, the general thing has
been, well, go for it if you can get the capitals
because the valuations are high enough and
enough capital you say, look, I’ll just run for it
and even if my valuation is above what I should
be, I’ll grow into it using the capital as a
way of making it happen. And that’s fine as long as
you actually, for example, on the pile of capital that
you’re currently raising, if winter comes you could
get to a break-even state because part of what
also frequently happens is that once you look badly capitalized, it’s very hard to raise additional money. So those are all the kinds
of factors that go into it. Now, for us, we kind of said, all right, each of our fundraises
was not actually trying to in fact manage as much capital
in the company as possible because we were steering
into a very different area than the other companies
that were fundamentally doing this massive capital
rise and running we’re going at because we had this notion
of professional identity and network is a platform. In our early stages,
like family and tribe, most of our competitors
thought it was companies were the primary thing, not individuals. We were the one doing individuals, so it was kind of like if they were right, doesn’t really matter if
we raised 2x the capital, or they raised 2x the capital. If we were right, again,
also doesn’t matter ’cause it was kind of a different thesis. And then when we got to the series D, which Bane led, we now view that it’s a market opportunity and we’re gonna go for it, but we’re really, we
were doing kind of more the reasonable scale blitzscaling versus ultra blitzscaling. We weren’t doing the hirer,
like hire people desperately. We were opening up a number of offices and we were beginning
to experiment with that and we were taking in
the communications burden that came with it, but we weren’t, like, for example, we
could say, well, you know, look, if it took an extra month
or two to make that higher, that’s not actually, in
fact, deathly for us. – Yeah. The only other comment
I would make is that sometimes when you’re raising here, you’re raising with different
uses of the capital in mind. So acquisitions is one that
we didn’t talk about here. It’s pretty common during
OS3 to actually acquire small companies, specifically
to fight over talent. – It’s a big layout for hires. – We didn’t put it in here,
we didn’t talk about it today because LinkedIn didn’t make
any acquisitions during OS3. (student asking question) – So, first question is, you know, when you’re in the
family and tribe stages, you hire people that can
scale the village and beyond or village and you know up, that you’re hiring deliberately in advance you’re hiring for it. And the second question is,
if you know that your hire, ’cause some of the people
you know, that you will, are very uncertain, do you
have the hard conversation? And the short answer is
there’s very few people that you actually have 100% confidence. It’s like a confidence interval. It’s very few that you go 95, 98, you know, like
this is gonna happen. When I look back at LinkedIn,
some people broke earlier than I expected and some people went much further than I expected. It happens that way ’cause it’s a new challenge, new circumstance, your own judgment. But part of that is then
what gives you an ability. If you have your conversation with people, where you say, look,
it’s not an entitlement, it’s not guaranteed you’re gonna be there, for example, part of the
way that I talk to people at LinkedIn is not I’m gonna
be the CEO all the time either. We are defined by our mission
about what we’re doing and how we’re being successful. And your job of me or your manager is to give you good feedback
and give you traction on that and if you start the
conversation gently early, you have a much higher likelihood of getting to a positive result. So if you essentially
go to someone and say, look, this particular part of your job, even something you particularly
love and hold on to, isn’t actually going to survive. Let’s have a conversation now. Go away, think about it for a couple days. Let’s have another conversation about it. While you’re having a
conversation, you’re articulating to the person, look, this is
what you’re really good at, this is why I really would
love it if you would stay here, this is why you still have a great mission if you’re doing these things. So you’re articulating
it, you’re not saying, oh you still can’t do that. You’re like, look, here are all the things that are really good. And ultimately, one of
the techniques that I use as a manager is I say
look, I’m not asking you in some metaphysical sense
to 100% agree with me in terms of what I’m
saying during judgment, but this is the way that we’re running. So it’s not like you
have to say, okay great, I’m really not very good at that. Fine, you can say I think I’m
better at that than you do and then just continuing playing and we’ll see how it plays out. So, those are the kinds
of things that play into those kinds of conversation. Oh, and then the very last piece of it is you’re almost never hiring, as part of not knowing the scale, you’re almost always hiring
a certain percentage, call it, you know, it could
be 50, it could be 70% of people who you know are
really good at this stage and that’s the stage you
should be solving for. Like they should be
100% good at this stage and you don’t know of the next and so you’re targeting
this stage, generally. – [Allen] So, let’s pause there because we’re a little over time. – Take one last question? – [Allen] One last question is great and then we also wanna talk
about what’s coming next. – [Voiceover] Thanks, can I (mumbles)? Well no, and so, what percentage of people
did you guys end up firing or like if you could give a range? Like was it highly uncommon or was it, you know, one every five people? – Well, generally speaking, we
would try as much as possible to manage people out through a process ’cause we thought that was the
right, kind of human thing, so technically it’s not a firing. ‘Cause that’s not the,
you know, here’s a paper, the security person’s
walking you to the door. (laughs) Right? You know, that kind of thing. So we’d had relatively
few firings, I think because we made a real
effort to try to do that. We did have a lot of people that maybe, probably, half per stage? Third per stage? – A third, I think probably a third. – Third? Okay, roughly, that we were
kind of going through that way. – [Voiceover] And that means
like, hey, you’re really good at these things, how about
you focus on those more? And then over time you end up– – Yeah, or like, hey
this isn’t really like, what you want isn’t really
gonna play out here. Let’s try to make it work so it works for both of you and for us. And I’ve had more conversations like that than I can remember the number. – Okay, so let’s talk
about what’s coming on the next two weeks. – [Reid] And by the way, we
can come back to questions on this all throughout the class. So, it’s not a speak once, forever– – And we’ll be around for a couple of minutes afterwards, for followups. So, Selina from SurveyMonkey
and Patrick from Stripe and Nirav from Nextdoor. I don’t know if you wanna
give an introduction to any of them at this point. You’re gonna meet them all
in the next couple of weeks. – So the high line is SurveyMonkey, most people who are
operative in the valley actually know it’s a
super interesting company that has a very
interesting business model. It’s somewhat stealth to everybody else. They may, you have taken the surveys, but don’t understand
how big the business is. Selina is the person who grew the product and engineering team for
doing it and it’s interesting. Patrick Collison is the CEO
and co-founder of Stripe, which is kind of the next
generation of payments. And Nirav Tolia is Nextdoor,
which is social networks for neighborhoods. And they’re all like
unicorn, unicorn, unicorn. Really. Right?
– [Allen] Yeah. (laughs) – [Reid] So.
– [Allen] All right, great. Thanks very much, we’ll
see you on Thursday.


Reader Comments

  1. Thank you Reid & Greylock! As a person who has failed numerous times at various stages of a project these lectures extremely accurate and detailed. Thank you for taking the time to share your insights, I'm sure many are going to benefit from these. It would be great if we could get access to the slides.

  2. I skimmed through as it got boring…they don't even mention Why and How They Scaled LinkedIn?…saw no architecture diagrams skimming through…just some shit about hiring more staff

Leave a Reply

Your email address will not be published. Required fields are marked *