Developing engineering leadership styles: conviction, policy and consensus.
Summary
TLDRIn this talk, Will Larson discusses the evolution of leadership styles for engineering leaders, emphasizing the importance of leading with conviction, consensus, and policy. He draws parallels with the culture shock curve, noting how some adapt to new leadership expectations while others struggle. Larson shares personal experiences and insights on developing these styles, suggesting they are learnable skills necessary for effective leadership in the current business climate.
Takeaways
- 🌟 Leadership styles evolve with changing circumstances, and effective engineering leaders need to adapt to these changes.
- 📈 The 'culture shock curve' can be applied to leadership, where some adapt and thrive while others struggle with new expectations.
- 🛠️ Three essential leadership styles for engineering leaders are leading with conviction, leading with consensus, and leading with policy.
- 🚀 Leading with conviction is crucial for making decisions in ambiguous situations where there is no clear consensus.
- 🔧 Leading with policy is necessary for maintaining consistency in decision-making as a company scales and requires standardized practices.
- 🤝 Leading with consensus is important when there is no clear leader or when a decision requires buy-in from multiple stakeholders.
- 🔄 The ability to switch between leadership styles is key to addressing different challenges and indicates a flexible and powerful leader.
- 📚 Leadership styles are learnable skills, not innate traits, and can be developed over time through experience and intentional practice.
- 📈 Examples of success and failure in leadership demonstrate the importance of adapting one's approach based on the situation rather than relying on a single style.
- 🔄 The speaker emphasizes the importance of being uncomfortable and stepping out of one's default leadership style to grow and improve as a leader.
Q & A
What are the three leadership styles discussed in the script?
-The three leadership styles discussed are leading with conviction, leading with consensus, and leading with policy.
What is the significance of the 'culture shock curve' mentioned in the script?
-The 'culture shock curve' is used as an analogy to describe how managers and leaders might experience a transition period when facing new or changing leadership styles and expectations, similar to how individuals adapt to new cultural environments.
How does the speaker describe the leadership style of the period between 2012 and 2020?
-The speaker describes the leadership style of that period as one where managers were expected to support their teams and promote career development, acting as an 'umbrella' to protect the team from unnecessary distractions.
What changes in leadership expectations did the speaker observe after 2020?
-After 2020, the speaker observed a shift in leadership expectations, with a move away from the supportive 'umbrella manager' style towards a need for leaders who can make tough decisions and navigate ambiguity.
What does 'leading with conviction' entail according to the script?
-'Leading with conviction' refers to making decisions and managing execution on critical or ambiguous problems when there's no consensus among stakeholders or when quick decisions are necessary.
Can you provide an example from the script where leading with conviction was applied?
-An example of leading with conviction is when the speaker had to address the perception of low engineering quality in a business unit at Carta. The speaker dug deep into the issue, tested hypotheses, and eventually concluded that the release philosophy needed to change to improve quality.
What is the role of 'leading with policy' in an organization?
-'Leading with policy' is about creating and enforcing consistent decision-making frameworks across a growing organization to ensure uniformity and scalability in operations.
How does the speaker define 'leading with consensus'?
-'Leading with consensus' is the process of building agreement among a group, especially when there is no clear executive direction or when decisions require broad agreement to be effective.
What is the speaker's view on the learnability of leadership styles?
-The speaker believes that all three leadership styles are learnable and not innate, and that even individuals who initially lack these styles can develop them over time.
What advice does the speaker give for developing leadership styles?
-The speaker advises reflecting on the leadership styles used in past problems, identifying the default style, and intentionally practicing other styles that are less comfortable but more appropriate for the situation at hand.
Outlines
🌐 Culture Shock and Leadership Styles
Will, the speaker, introduces the topic of leadership styles in engineering, drawing a parallel with the culture shock curve experienced by JET program teachers in Japan. He discusses how the expectations and styles of leadership have evolved, especially post-2020, and suggests that effective engineering leaders need to master three distinct leadership styles: leading with conviction, consensus, and policy. He also shares his personal journey from teaching English in Japan to becoming a CTO and author, emphasizing the importance of adapting to new cultural and leadership landscapes.
🛠 Leading with Conviction and Policy
Will delves into the first leadership style, leading with conviction, which involves making decisive calls on critical or ambiguous issues. He argues that this style was often deemed inappropriate in the past decade but has become essential in recent years. He contrasts this with leading with policy, which is crucial for maintaining consistency in decision-making as a company scales. Will shares his experiences, highlighting the importance of informed decision-making, testing hypotheses, and pushing through execution despite friction. He also discusses the challenges of implementing new programming languages and the need for clear communication and enforcement mechanisms in policy-led decisions.
🤝 Building Consensus in Leadership
The third leadership style discussed is leading with consensus, which is necessary when there's no clear executive direction or when decisions require broad agreement. Will outlines the process of identifying a decision that could accelerate progress, evaluating the feasibility of reaching consensus, and forming a decision-making group if needed. He shares a successful case from his time at Uber, where he facilitated service decomposition without a top-down mandate, and a contrasting experience at Comm where a lack of consensus during layoffs led to inaction and reliance on the CEO for resolution. Will emphasizes the importance of aligning incentives and the challenges of building consensus when stakeholders have divergent interests.
🔄 Adapting Leadership Styles for Effective Problem Solving
In the final paragraph, Will advises on how to identify and adapt one's leadership style to suit different problems. He suggests reflecting on recent challenges and noting which leadership style was applied. He encourages leaders to step out of their comfort zones and try different approaches to expand their capabilities. Will's advice is to be mindful of one's default style and to intentionally practice other styles to become a more versatile and effective leader. He concludes by thanking the audience for their attention and highlighting the importance of continuous learning and adaptation in leadership.
Mindmap
Keywords
💡Leadership Styles
💡Conviction
💡Consensus
💡Policy
💡Culture Shock Curve
💡Adaptation
💡Engineering Leadership
💡Hypothesis Testing
💡Enforcement Mechanisms
💡Incentives Alignment
💡Executive Escalation
Highlights
Discussion on three essential leadership styles for effective engineering leadership: leading with conviction, consensus, and policy.
The importance of adapting leadership styles in response to changing business environments and expectations.
Analogy of the culture shock curve to the shifts in leadership expectations over time.
The necessity of leading with conviction in situations where there is ambiguity and stakeholder misalignment.
Leading with policy as a strategy for consistent decision-making across a growing organization.
The role of leading with consensus when executive leadership is not engaged enough to drive decisions.
Example of successfully leading with conviction by addressing quality issues in engineering work.
Case study of failure in leading with conviction due to lack of follow-through and enforcement.
Implementation of a successful policy through candidate review at Stripe, highlighting the importance of enforcement mechanisms.
Lessons from the failure of agile policy implementation at Stripe due to lack of follow-up and training.
Building consensus at Uber through service decomposition without top-down mandates by aligning incentives.
Challenges in achieving consensus during layoffs at Comm due to disagreements on budget cuts.
Advice on identifying and developing the three leadership styles by reflecting on past problem-solving approaches.
Encouragement to step out of leadership comfort zones to develop a broader range of leadership skills.
Final thoughts on the importance of adapting to new leadership styles and the learnability of these styles.
Transcripts
hey folks this is will I'm here to talk
a little bit about leadership styles and
particularly three that I think are
really important to being an effective
engineering leader kind leading with
conviction leading with consensus and
leading with
policy CTO aarta prior of the CTO atom
done a bunch of other things before that
and also written a handful of books most
recently and during Executives primary
worth wor
3 and the first job I had out of college
was you know not not writing software
but actually teaching English in Japan
as part of the the jet program so
Japanese exchange teaching basically
take a bunch of 21y olds um throw them
into the classroom in Japan to teach
English and know English speaking
culture sometimes from New Zealand or
Australia the US the UK you know
whatnot one of the things they taught us
was this culture shock curve and so you
know most of the people that were
teaching in Japan had never never been
outside of the country before just like
short they' certainly never been working
outside the country before and so they
didn't know what to expect and this
helped us understand kind of the the
year we were in for and so there's this
honeymoon period you feel really really
good and then after you know two three
months often people started to feel
really bad they they were upset about
everything that was different around
them and wasn't like where they grew up
wasn't what they were used to and they
were really
frustrated then you know slowly but
Surly kind of dig out he start to adjust
and finally get this adaptation stage
where you're basically as happy as you
were before before you came to new
culture but what I noticed is like a lot
of folks went through this and had a
great experience but some folks did not
ever get out of kind of the bottom of
the curve some people hated it and were
frustrated and just stayed there those
those were those were not happy
campers and I've been thinking a lot
about that culture shop curve as I've
been talking to more and more managers
and you know leading myself over the
past four years I think for for someone
who came up as a manager in that 2012 to
to 2020
period things were feeling pretty good
for a lot of that and part of that was
like the zero interest rate policy that
was putting more and more money into
startups where there was like less and
less friction around like how do we get
enough money hey just just raise more
it's
easy um but also there's like this real
idea that the goal of managers in that
era was to support their team to promote
Career Development how to get your team
EXC excited and happy there's this idea
of kind of the the umbrella manager
and your your job was to protect the
team from the company distracting the
team from doing useful things and that
was kind of what the the era was and hey
you hit 2020 and and things have really
shifted since um first you know there
was like the boom of over hiring then
you know kind of the the layoffs to kind
of reduce expenses back um but even now
things are things are just pretty
different and the expectations of what
we want from leadership the different
leadership styles that we think are
representative of a great leader he are
really shifted and similar to the
culture shock curve that I saw in the
jet program some people have been like
oh man I love the old model had a really
rough time then popped out kind of
figured out that the new way to succeed
some people are still pissed and um you
know convinced that the the last decade
was the right decade the new decade is
just misguided I want to talk a little
bit about um what that means and and how
to how like move forward if someone that
is you work with is is stuck there or
maybe that you in some ways feel a
little bit stuck yourself and I feel
like this new moment is is like bad and
they we used to have like good
leadership and now it's an era of bad
leadership I've heard that a lot
although not everyone wants to kind of
say it say it out
loud so the core hypothesis I have is
that for folks who are stuck in the
bottom of that um most of them are
missing at least one core leadership
style
and I think there's three leadership
styles you need to be pretty effective
as an Eng leader need to lead with
conviction lead with consensus and then
lead with with policy and I'll talk
through all
three also the corollary I think these
things are are learnable not inate and I
think it's easy to convince yourself
that these are innate to who you are you
can't learn that this conviction style
because it feels alien or wrong or
consensus is is slow and dumb you don't
need to learn that but hey I actually
think all these are learn able and as
someone who came in with like none of
these leadership styles have been able
to systematically build them over time
think you can too and so can someone on
your team who he's been struggling with
these so I want to talk a little bit
about these three Styles give some
examples of success and failure that
I've personally run into with all three
Styles and talk about how to develop
them if you or someone you work with is
missing one of the the
three so there are basically three
different styles here right top to
bottom so leading with conviction and
this is when you personally decide to um
make the decisions and and manage
execution on these critical or ambiguous
problems I think for the past decade you
were told that this sort of leadership
is bad leadership but I I actually think
it's it's a necessary skill and I've
been using more of it over the last four
years than in the prior um decade
combined and really important for
decisions when there's just no way to
use the other the other kind of
approaches so deciding to add new
programming languages you're going to
really piss off um you know your depth
tooling team or your infrastructure team
but hey maybe you just did like a new
acquisition who only works in that new
language and there's no way to kind of
get to the bottom of that other than
building um an informed perspective
yourself and making making that
choice then leading with policy
particularly relevant when your company
is growing quickly you're trying to go
from a spot where you can just rely on
one or two people who have all the
context making great decisions to
figuring out how you consistently make
the same decisions across a much larger
population of folks than I need to make
choices think quarterly planning hiring
um promotions these are all cases that
this kind of leading with policy comes
up quite quite a bit and finally leading
with consensus I think as Leaders we we
generally don't want to put our teams in
spots where they have to lead with
consensus but often we um have to look
around and say like hey you report to
the CEO or you report to the CTO and
they're not necessarily going to lead
with conviction on every every problem
they just don't have enough time to even
if you want them to and sometimes that
are going to put you in spots where you
have to make the choice of either not
making progress at all or building
consensus with a group who maybe really
doesn't want to build consensus um and
you have to be able to do that otherwise
you're just you're just dependent on the
the executive to to do it and then
you're not really a leader you are just
an extension of another leader which you
know get pretty far with but again to be
a flexible um powerful leader that can
work in many different situations need
to have this one as well
so leading with conviction um this is
when um there's you know stakeholders
that can't possibly align there's a ton
of ambiguity and people will just get
stuck and the the pattern here is
basically to go really deep on the
context to educate yourself maybe
talking to customers and internal
stakeholders who who know a lot test the
decision widely and test both with like
peers outside of the company tests with
stakeholders internally and also find
ways to actually like try the the
decision and see if it Works um then
communicate what decided and then push
push push through the friction and
anytime you're leading with conviction
there's going to be a reason you're
doing it versus something else and it's
usually that there's going to be
friction so you have to get comfortable
pushing through even when it's it's a
little bit
messy so I think a great example of
leading with conviction is some work
that I've been doing recently with the
team at carda and there was one of our
business units where there's a
perception that the quality of the
engineering work we were doing wasn't
that high
and so there there's all these easy
narratives that come out when quality is
not high it's like okayy the team
doesn't care about the work we're doing
it's like oh the team is sloppy oh the
product managers aren't giving us clear
enough specifications for for the the
work to be done so you start with like
all these different ideas and and
there's really in this case that there
was no way to like figure out what to do
um no no one would have no one even
agreed on the diagnosis of what was
going wrong let alone the solution to it
um and so as we started kind of trying
to understand this perception that the
quality wasn't that high on some of our
work um started digging in one of the
first ones was okay maybe we don't know
how to test our code and so one of the
first hypotheses is like we need to roll
out more uniform testing standards and
and as we kind of explored that idea we
did find that um on the front end our
our testing kind of patterns were were a
little bit like underdeveloped and so we
could do a little bit more work there
but it really didn't explain the general
observation and so so we had to keep
finding like thesis and then like
testing them and then like you know a
lot of those thesis kept falling away
and where we ended up um which is not at
all where I expected to end up was this
observation that um our biggest problem
was was likely our release philosophy
and so we were releasing to um fun
venture capital and private Equity Funds
who are all um extremely heterogeneous
they have almost nothing in common
across the entire population little
pockets do and as we tried to release to
all of them we we would consistently run
into incredibly nuanced errors um so we
had to rethink how we deployed and new
functionality and kind of deploying to a
small cohort validating eror and then
cohort by cohort building up to the
entire population over time and it
wasn't that we were sloppy it wasn't
that our software didn't work it's said
it worked really well for certain
Pockets but not the entirety and the
there was no like uniform consistent
kind of thing and need could solve we
had to like roll pocket by pocket out to
validate expand across relatively
homogeneous groups within the the
broader heterogeneous population and
this has really impacted our our quality
and we would have never gotten to this
conclusion if we had not had the kind of
desire to dig in deep and continuing to
iterate through hypotheses even as some
of them were
wrong and so this is interesting because
I I tried to do something pretty similar
at my my last job at at com and it
didn't go very well and so it's
interesting to compare and contrast so
as I went into CM I realized there was
no agreement about how we should test
functionality this was leading to some
bugs kind of escaping through and I I
had like a clear point of view about
where we did which types of tests um
pulled together a group gave them my my
desired approach tasked them with it
came back a couple months later and
nothing had happened kind of the
different stakeholders of the group
disagreed and so they wrote like a memo
that was extremely um you know lowest
Comm nominator didn't really change
anything
thing and it never went anywhere and so
the you know the the comparison point is
I had a clear thesis I handed it down to
the team but I didn't actually put
myself in that weekly meeting where
we're like driving execution looking for
disagreement finding kind of the blogers
and pushing through and so that success
and and failure in this case really came
down to not just having the big brain
idea but being in the the weekly
meetings where we actually got from big
idea to the details coming together and
I think what I particular in the card
example it's that the big idea was you
know inevitably wrong and it's only got
refined into something good in the
weekly work that that we went through so
second um kind of approach leading with
policy and this is again when you're
trying to get um many different
individuals to make generally the same
type of decision across your
organization um and in the process here
you know look at a recurring decision
that's happening a lot already document
How It's Made um roll out the policy to
folks and then crucially figure out the
enforcement mechanisms that actually
caused the policy to get used rather
than just be a document that kind of
sits
there so I think a success story on this
is candidate review at stripe so when
stripe was small like every hiring
decision was getting made by by Patrick
the CEO then as things scaled a little
bit we we Ste continue to have this
process where the CEO made every
highering decision but eventually just
got to be too much and so we roll that
candidate review for engineering and a
few different functions but within
engineering basically how it worked like
every every week we had three meetings
had a group of like four orifi of
rotating folks including like a senior
leader or CTO and one of our our senior
Recruiters on engineering plus like two
or three other kind of midlevel
engineering leaders as well and we just
talk through each of the packets and
we'd give feedback and we'd kind of
approve thumbs up thumbs down or hey you
need to to provide additional context on
it and we always had some who you know
two two people really who could escalate
decisions when something went a little
bit wrong with the group so if if thing
we got to the seemingly wrong decision
either the senior engineering leader
reporting the CTO or the recruiter or or
or both if they disagreed on a given
topic could escalate up their reporting
chains make sure that we got a little
bit more like oversight into the
decision this wasn't perfect but it
worked it worked pretty
well um then looking at the failure mode
at stripe again um we um want the engine
engering leads um decided to roll it
agile got kind of this mandate from from
the head of engineering at the time um
and you know they wrote up a document
communicated to all of engineering that
we were an agile shop capital
A and and that was it and everyone at
this meeting was like a little bit
confused like hey I guess we can do
Agile and and that was the end of it
there was nothing more there was no
followup there was no review there was
no discussion there was no training and
so this went nowhere and the policy was
actually total totally reasonable but
the the real difference between these
two is like the first one we had like
these enforcement mechanisms where we
actually looked at you know week by week
was it getting implemented had
experienced folks kind of helping Drive
consistency of the approach and we had
escalation mechanisms for when we got it
wrong to to get to the right people to
come in and inspect the the error and
and make sure it got fixed second one
good document but but no follow through
no enforcement mechanism so it didn't
actually anywhere pure pure
Optics then finally like working with
consensus so this is really useful when
there there you know the first example
there's you know you have this um
engaged executive you can actually like
drive a decision sometimes you need to
make a decision but there is no engaged
executive to actually do it what do you
do then and it's it's developing
consensus right and so the The Playbook
here you look for a missing decision
that would really facilitate moving
faster then you try to decide if you can
just not do this work because consensus
is a little bit slow takes a fair amount
of energy to to build you really look
for the decision maker if if you can't
um if you can't convince yourself to to
skip making the decision and then if you
just can't find the decision maker or
they're not engaged can't prioritize the
work then you form a decision- making
group and Lead that through and I you
know don't love this but it's it's been
most of the successful work I've done on
an executive teams has really been
consensus building and leading through
consensus if the CEO had a clear point
of view on a given decision they would
have already told you what they wanted
and driven the consensus that way or
driven conviction to a decision that way
but in this case they they clearly don't
think it's that important relative their
other work so they're not working on it
the only way then for you to make
progress on it is is through consensus
so again couple couple examples I think
one decomposition in Uber is still one
of the things I'm I'm proudest of that
I've worked on and in part I'm really
proud of it because we had no top- down
mandate to to do this There Was You Know
General support for it but hey there
wasn't this perspective that
decomposition was obviously going to be
the path forward and we couldn't force
people to move out of their their
current Services um we could just like
tell them what was there and so instead
of looking at this hey we're going to
force you we looked at like what could
we do to incentivize people to kind of
get aligned with us we made provisioning
Services really really really easy and
then we educated folks like even in
onboarding like each person who came
into Uber and Engineering did a service
provisioning just to see how easy it
was and then we we got people to opt in
and we couldn't force them to but we we
could make it so easy that people came
to us um eventually there was a mandate
from one of the engineering directors um
but again we we didn't force him to do
it and we couldn't force him to do it we
just made it so obvious to him that this
was an easy path forward for him and his
team that he made the decision kind of
drove the the the customers to us and we
were able the change I think one of the
the most important parts of the
company's architecture without any sort
of top down mandate which is which is
pretty
remarkable then then the flip side of
consensus was we we did a a layoff at
Comm and we you know as an executive
team couldn't agree on where the cuts
needed to come from and so the the chief
product officer and I had a perspective
that we should start from looking at
comparable companies looking by function
like GNA SNM R&D and the the growth rate
for our company and what are the Target
spends we should have and then use that
to kind of anchor each of those
functions to the to the right amount of
spend and then cut within those
functions to get there um but you know
some of the functions that were a little
bit outsized in their spend um didn't
didn't like that idea and so we we
weren't actually able to get consensus
and so we ended up with a spot where
people were just kind of invisibly
escalating to the CEO instead of
actually the having the executive team
work together to get the business to to
the right place and so this this was
like a little bit I think a a lost
opportunity where if we could have just
aligned a little bit more on what we
were trying to do I think we could have
come up with a much better series of
decisions instead it was just you know
little bit of a of a
mess and again looking at these two
different differences I think on the
first one we really focused on getting
like the the clean path to align
incentives of the different stakeholders
we we couldn't we couldn't force it but
if we made the incentiv so easy um that
people come along with us the second one
we we just couldn't get the incentives
to work and so ultimately we we weren't
able to drive any sort of align decision
we just kind of got stuck and then had
to wait for the CEO to kind of dig us
out of of the
mess and so looking at these three
different styles um if you want to
figure out if you're using them like the
biggest advice I have to you or to
someone on your team that could work on
this is to just look at um the next five
hard problems you work on and write down
like what style did you actually use to
work on them maybe consensus maybe
working with conviction maybe maybe
policy but just write down like what is
like your your go-to like strategy for
these types of
problems and then you know once a month
um when you take up a problem don't use
your default and and try to use the
style that's actually appropriate to the
problem but if if you're so comfortable
with one pro uh approach you might just
lean on it for everything and the key
thing is pick something uncomfortable
sometimes s you cannot let yourself just
do what's comfortable you'll continue to
be you know stuck in in that kind of
policy and that that leadership style
rather than expanding what what you're
capable of doing over time it would get
a lot better and that's it so really
appreciate yall um listening and thank
you so much
Посмотреть больше похожих видео
5.0 / 5 (0 votes)