00:02
hi everybody I’m Josh constant here at
00:05
TechCrunch Disrupt SF I’m here with
00:07
Peter Thiel of founders fund and Metro
00:09
capital now you’ve talked a lot about
00:11
how you need to choose companies that
00:13
might seem crazy at first to the average
00:15
person otherwise it would be an idea
00:16
that somebody already taken how do you
00:18
tell the difference between crazy and
00:20
crazy like a fox you’re something that
00:21
will actually pan out well you try to
00:24
you try to just always evaluate the
00:27
substance afresh so does the technology
00:30
work you can ask questions about what
00:33
what’s the prehistory of the founders
00:35
have they been working together for a
00:36
while are they are they going to give up
00:39
at the first sign of trouble so I think
00:41
you you want to always focus on the
00:43
substance there are no shortcuts so you
00:46
try to avoid you know coming in with
00:48
biases about what can and can’t work
00:49
ahead of time yet there’s always a
00:51
temptation to come with all sorts of
00:53
sort of one-sentence rules and we’re as
00:56
guilty of that as anybody but but I
00:59
think what’s what’s gone best is you
01:00
find something interesting about it and
01:02
then you try not to get to a yes or no
01:04
decision right away you try to really
01:06
understand the substance keep an open
01:08
mind and understand that so substance
01:11
over-over process every time so it’s
01:14
surely a hard question to answer because
01:16
it’s very generic and everyone every
01:18
startup is different but when you hear
01:19
an idea that you’re wondering if it can
01:21
really work what is it your go-to
01:23
follow-up question to ask an
01:25
entrepreneur pitching you about and what
01:27
is their their philosophy or their
01:29
long-term vision and how likely they are
01:31
to succeed at it well well there’s sort
01:34
of you always want to get the team the
01:36
technology and the business strategy so
01:38
you have to get all three of those to
01:40
work and so if people are if you want to
01:42
talk about the technology we’ll talk
01:44
about the people in the business
01:45
strategy so you go to you go to the the
01:47
other two topics if they don’t want to
01:48
talk about them that much so I wanted to
01:51
ask you a little bit about your
01:52
philosophy and the philosophy of the new
01:54
rich that’s come out of the technology
01:56
scene you know you’re known for having
01:57
very strong perspectives I personally
01:59
went to the seasteading festival of
02:01
ephemeral really fascinated by that way
02:04
of looking at things where we can solve
02:05
the problems ourselves we don’t
02:06
necessarily have to wait for somebody
02:07
else what do you think do you think that
02:09
there’s a different value system amongst
02:12
people who are becoming wealthy out of
02:14
the technology scene opposed to you know
02:16
industries of old it’s always hard to
02:19
generalize but it I I do think there’s
02:21
play something different whether you
02:22
made your money in computers or in say
02:25
resources in the Congo or something like
02:27
that so I do think I do think that
02:29
there’s something generally good about
02:31
technology and generally very positive
02:34
about this as an industry I’m obviously
02:35
a bit biased but I do really I do really
02:38
believe that I think I think
02:40
philanthropy is an area that deserves to
02:43
also be rethought and I always like
02:45
asking these contrarian questions and so
02:46
if the contrary in question businesses
02:48
what great companies nobody’s starting
02:50
in on implants would be the contrarian
02:53
question I think is always something
02:54
like what great cause does nobody want
02:57
to support and and so I always uh I one
03:01
of the questions I always like to ask is
03:02
why is it cause unpopular I don’t want
03:04
to give money to popular cause I feel
03:05
those are well enough funded I think the
03:07
unpopular causes that deserve to get
03:09
more its me it seems like when people
03:11
come up so quickly in the technology
03:13
world because of the way that you know
03:14
equity structures work and how quickly
03:16
you can get to market and make something
03:18
really big with the modern distribution
03:20
channels that you know if somebody
03:22
becomes rich in five years it feels like
03:24
maybe that money doesn’t really belong
03:25
all in my bank account like it’s kind of
03:27
belongs to the world it was kind of
03:28
random that it ended up with me opposed
03:30
to somebody who maybe came up through a
03:31
more hierarchical world worked every day
03:33
for twenty years and when they get to
03:34
the end of it they say this money really
03:36
is mine do you see any of that
03:37
perspective that people feel like they
03:39
don’t really have a right to the money
03:40
that they’ve earned in technology and
03:42
that it’s better for them to give it
03:43
away or like what is the philosophy of
03:45
people who are taking the giving pledge
03:46
or making those really big donations
03:48
well I think I think in general it is um
03:52
it is that that they they want to give
03:55
something back to society I’m not sure
03:57
they I think that mostly do think that
03:59
they actually deserve all of it but I
04:01
think these are some of what they got I
think the reality is that what’s unusual
about the tech industry in Silicon
Valley
is that the inventors are capturing
anything at all the history of
innovation has been one where most the
people who invent things get nothing at
all and so the Wright brothers came up
with the first airplane but didn’t get
rich or you know even the original
Edison versus Tesla you say Tesla is the
greater in
and that was the better way to go but
somehow Edison edged out Tesla and so
most of innovation has actually not not
gone to the people who came up with
things you have to to make money you
have to do two things number one you
have to create something of value for
the world and number two you have to
capture some fraction of the value you
create and and often people have
completely failed of doing the second so
I think Silicon Valley is very unusual
04:51
in that there’s this large class of
04:54
innovative people that are actually able
04:56
to capture some fraction of the value
04:58
they’re creating that’s great so now
05:00
that we have founders who are making a
05:01
fortune by making something everyone in
05:03
the world wants hopefully they can make
05:05
some difference for the better in the
05:06
world as well yes yes and I always I
05:09
always would say that you want to make a
05:11
difference for the better in a way
05:12
that’s not just looking for status not
05:14
just in a respectable way but people
05:17
should try to be courageous in their
05:19
philanthropy
I never planned to do a series of articles, but this write-up almost came as a natural follow-up to the post I wrote last week about Context over Control and the future of remote work. Last week I explored the ground rules of the future of work. Specifically, I talked about how and why the fundamental prerequisites of work have changed. Here I’m taking a step further and describing some of the literature behind today’s organizations’ communication systems. What they’re made of, why they can be so impactful in today’s organizations, and how they’re related to the concept of remote working.
On to the write-up.
A good internal design communication system is one of the most important leverages an organization can have to make an impact.
In 2011, a post came out under the name of Stevey’s Google Platforms Rant. I have revisited the article over and over again over the years (if you haven’t read it yet it really is fantastic) and to this day I think this is the single best article I’ve ever read about organization architecture and the management of IT.
Yegge’s rant is about what he’s noticed after spending 6 years at Amazon and 6 years at Google.
In short: Amazon figured out how to serve customers. It started out serving Amazon.com but critically, as Yegge noted, the relationship was very much at arm’s length: AWS has always treated Amazon.com no differently than any outside user, which prepared it to offer the right sort of services for any scalable web service. On the other side, Google, with its Cloud provider, Google Platform, does not serve most Google Products. In other words, while Amazon did manage to have discipline towards internal dogfooding, Google didn’t. And that made a world of difference.
In recent years, I started to realize that what I noted above was actually the gist of that post, but it was not the most interesting thing about it.
In the article, Yegge (again, read it if you haven’t yet) describes an internal Email Jeff Bezos sent to the 150-odd Amazon employees. Quoting Yegge’s post, the email was along these lines:
1) All teams will henceforth expose their data and functionality through service interfaces.
2) Teams must communicate with each other through these interfaces.
3) There will be no other form of interprocess communication allowed: no direct linking, no direct reads of another team’s data store, no shared-memory model, no back-doors whatsoever. The only communication allowed is via service interface calls over the network.
4) It doesn’t matter what technology they use. HTTP, Corba, Pubsub, custom protocols — doesn’t matter. Bezos doesn’t care.
5) All service interfaces, without exception, must be designed from the ground up to be externalizable. That is to say, the team must plan and design to be able to expose the interface to developers in the outside world. No exceptions.
6) Anyone who doesn’t do this will be fired.
7) Thank you; have a nice day!
Now I think that this internal email is what has actually stuck with me the most. Bezos realized that he had to change the internal communication infrastructure before he could actually change the byproduct of the organization.
He understood that a radical organizational change was required to arrange the internal dynamics in a way that would allow the creation of something like AWS.
Specifically, what he got right was an internal communication system designed to (1) embrace accessibility as its most important commandment in order to (2) enable a strong platform mindset and (3) incentivize extreme dogfooding.
While the third point makes all the difference in the world, what Amazon really did get right that Google didn’t was an internal communication system designed to make all the rest possible.
Having teams acting like individual APIs and interacting with one another through interfaces over the network was the catalyst of a series of consequent actions that eventually made possible the realization of AWS in a way that couldn’t have been possible otherwise.
To this day I think the Amazon example might be one of the clearest case manifestations of Conway’s Law.
Organizations which design systems are restricted to producing designs which are copies of their own communication structures.
Frameworks for organization structures
Back to the initial premise of this article, it’s amazing how powerful and impactful internal communication infrastructures can be to an organization: from day-to-day operations, to its culture, to the actual end-to-end user’s product.
There is one more additional point about Amazon’s case that is worth keeping in mind: Amazon didn’t perform any reorganization after Bezos’s mandate. The internal organization structure was already there: same people, same teams, same chain management. What changed was the airflow: the policies, the processes, and the communications interfaces that regulated the internal dynamics of day-to-day operations. In other words, Amazon’s highly divisional organization was already suitable for such a big change.
Communication infrastructures are highly dependent on the organizational structures that are already in place. Because organizational structures depend on the people who are part of the company, these changes tend to happen over long time-frames (if they happen at all).
Lately, I’ve been thinking about what possible types of organizational structure there are, and what they empower.
Functional organizations
Functional organizations are organized around areas of expertise. Apple might be one of the most renowned examples of a functional organization. In the case of Apple, that means that design is one group (previously under Ive), product marketing is another (under Schiller), and operations a third (under Williams, who is also Chief Operating Officer), etc.
Functional Organization structure came to light as the embodiment of Fayol’s idea of unity of direction. The consequence of this principle is encapsulated in the name itself; an optimal shared level of consistency, coordination, and alignment across the entire organization. This, as in the case of Apple, often translates to vastly superior customer experiences.
On the other end, functional organizations tend to come with a lot of internal bureaucracy. Deep alignment and consistency come at the cost of a slower and more rigid organization. Because of the low level of team autonomy, in these types of organizations, innovation tends to have a top-down (rather than bottom-up) trajectory.
Divisional organizations
At the perennial opposite of the spectrum, we find divisional organizations. Divisional organizations sacrifice Fayol’s principle of unity of direction and are organized around products rather than expertise. They started to be part of the playbook after DuPont’s famous reorganization in 1900.
Amazon is probably the best-known case of divisional organization. The two-pizza teams (introduced in its early years, and still used today), their most common internal division, are the ultimate embodiment of a divisional organization. In these organizations, product teams act as independent entities within the wider organization. They have their own marketing, sales, engineering and finance functions so that each has autonomy and accountability. Each product has its own profit-and-loss statement (P&L), whereas Apple famously had a single P&L.
Two-pizza teams excel at agility, structure, clarity, speed of (mainly bottom-up) innovation, meaning, and impact. On the other hand, this profound divisional organization comes at the cost of inconsistency, a high-maintenance communication structure and an inferior product experience. As Yegge rightly noticed in his rant:
“Amazon’s recruiting process is fundamentally flawed by having teams hire for themselves, so their hiring bar is incredibly inconsistent across teams, despite various efforts they’ve made to level it out.”
Hybrid strategy
Then there are hybrid organization structures. Fundamentally, these are either functional organizations that adopted some divisional principles or vice-versa. One example that immediately comes to mind is Netflix.
“As companies grow, they often become highly centralized and inflexible. Symptoms include:
1) Senior management is involved in many small decisions.
2) There are numerous cross-departmental buy-in meetings to socialize tactics.
3) Pleasing other internal groups takes precedence over pleasing customers
4) The organization is highly coordinated and less prone to error, but slow and frustrating.
We avoid this by being highly aligned and loosely coupled. We spend lots of time debating strategy together, and then trust each other to execute on tactics without prior approvals. Often, two groups working on the same goals won’t know of or have approval over their peers’ activities. If later the activities don’t seem right, we have a candid discussion. We may find that the strategy was too vague or the tactics were not aligned with the agreed strategy. And we discuss generally how we can do better in the future.
The success of a “Highly Aligned, Loosely Coupled” work environment is dependent upon the collaborative efforts of high-performance individuals and effective context. Ultimately, the end goal is to grow the business for bigger impact while increasing flexibility and agility. We seek to be big, fast and nimble.”
Frameworks for internal communications
Internal communication frameworks are like air traffic control, coordinating internal policies, procedures, and interfaces. There are many ways communication frameworks can be described, but generally, they can be part of a single spectrum: synchronous communication vs asynchronous communication.
Synchronous vs Asynchronous communication
In traditional, co-located teams, most of the communications take place here and now, or synchronously. Everybody on the team is focused on the same things at the same time. Meetings, brainstorming sessions, 1:1, sync-ups, discussing problems over lunch, catching up in coffee-breaks, and many other ways to interact with each other are all synchronous forms of communication.
These dynamics change when physical proximity is not part of the equation anymore. Remote teams tend to have little synchronous time and most of the work is coordinated online and asynchronously. Asynchronous frameworks allow us to exchange information at a convenient time for each participant in the process, independently of each other. Data is sent and received with a delay.
Interestingly, communication frameworks can work quite independently from the organization’s location. I haven’t found a single chart that captures the essence of how current internal comms frameworks are related to companies’ locations:
Co-located synchronous
This is by far the most traditional and common approach of the last century. Apple is one of the major embodiments of a co-located synchronous organization. Most of the companies that follow a similar configuration tend to be highly functionally organized.
Back to my previous point: these organizations are highly coordinated and synchronized through centralization, but they tend to be rigid and slow. Moreover, this slowness increases exponentially with headcount, making reorganization even more difficult.
There is one more additional point about co-located synchronous organization that is worth keeping in mind: it’s hard for them to implement principles of divisional organization and asynchronous communication, or to embrace a remote strategy.
Co-located asynchronous
These are co-located companies that operate from the same location with asynchronous principles, Amazon being a prime example.
To return to Yegge’s rent, the entire communication system in Amazon privileges autonomy and independence. SOA-mindset is deeply embedded in the culture. Teams are used to interacting like APIs in an asynchronous fashion.
No doubt meetings are quite limited at Amazon as they represent an interference in the usual communication between independent units or teams, as opposed to Apple, where meetings are catalysts of ideas and discussions.
Distributed asynchronous
At the very end of the spectrum, an organization can also be distributed. In distributed organizations, there’s no physical presence (no HQs or other forms of operative locations). Teams’ autonomy is highly decentralized and localized at the edges. The asynchronous approach is the de facto setting for these types of configurations. It’s common for distributed teams to rely on divisional organizations and turn physical constraints into a competitive advantage.
Distributed synchronous
Co-located organizations do not necessarily operate according to synchronous principles, similarily distributed organizations do not necessarily have to operate in an asynchronous fashion. InVision is one of the biggest organizations out there that embraced this hybrid form of collaboration early on. As a distributed organization they don’t have any HQs but they tend to operate from the same time zone (9 AM – 9 PM EST).
Interestingly, they’ve decided to settle on a particular compromise. Fundamentally, they are still a synchronous company but with a 50% time span (and some policies surrounding that) , and a minimum of 4 hours overlap. They still need to be async on any given level, but that minimum sync time gives them room to be less rigid around documentations and be more responsive in case of emergency.
Remote asynchronous
Asynchronous frameworks for internal communication give organizations that have a physical presence (main HQ) the opportunity to expand and embrace remote working more easily. Basecamp is headquartered in Chicago, even though a huge part of the team is spread out across 32 cities around the world.
Asynchronous = Optionality
As I hope is now obvious, there’s no single way to organize remote work and there’s no evidence to form any strong conclusions about the efficacy of one configuration versus another.
One thing that is worth considering though is that an asynchronous setting gives more optionality. It allows you to reorganize the company in a divisional organization more easily and embrace remote working even if you’re co-located. Everything that works in an async fashion can also work sync but not vice-versa.
On the other hand, it’s way harder for co-located sync organizations to expand into adjacent areas and experiment with other configurations.
As more and more teams are embracing some form of distributed model because it widens the talent pool from which they can recruit, transitions to asynchronous communication will become more evident.
The caveat here is that asynchronous communication can be very daunting at the beginning because it requires more processes, documentation, and infrastructure than synchronous communication. While synchronous communication tends to scale linearly, asynchronous communication follows a logarithmic trajectory and its efficiency surplus is only observable in the long run.
Conclusions
In my latest essay, I argued that part of the neoclassical economic literature on marginal productivity is fundamentally flawed. And because it’s incredibly hard to measure software productivity, it’s also quite impossible to hash out what can be the most productive remote configurations. That being said, here are some final thoughts based on these considerations:
An organization’s communication system can be one of the most important leverages you can have to make an impact on productivity. Be very intentional about it.
There are different distribution patterns for teams, not just a simple co-located versus distributed dichotomy. The advantages, disadvantages and effective techniques for multi-site teams will often differ.
Most groups of people will be more effective when working co-located due to the richer communications they have. But don’t forget that some people seem to be more effective in a remote-first model.
Never forget that organizations can adopt asynchronous communication or synchronous communication regardless of their physical presence.
Asynchronous communication is the form of internal communications that gives most optionality, and it’s preferable if you’re going to opt for a divisional organization. But it comes at the cost of high maintenance, which can be daunting, especially in the early days.
When using a remote working pattern, pay attention to how the communication patterns form. Invest in improving communication, including travel and technology.
