♪ [music] ♪ - [Tamao] So, thank you for coming. We're so excited. - [Kelsey] Was it optional? - Absolutely was. I was so excited. Well, I was very excited to hear, first of
all that...I mean, even at a Google event, people were just saying, "It's so hard
to get Kelsey to speak." And I was like... He's at DevRelCon. - Aces. - Yes, exactly. And I was excited that you particularly
cared about this topic, the one that I was sort of putting
out there as the theme. So, we'll get to that. And I'm really excited to hear all your
thoughts and all your great experience. So, as you mentioned,
as Matthew mentioned, so, we're in the Kubernetes
space in our world, and there are literally people
from Google who bought tickets just because they're like, "Oh,
Kelsey is speaking." So I'm really excited. So, for those of you who may not be in our
space, yeah, tell us a little bit about your path and what you do,
how you describe it. - Right. So, this is my very first time
being in a DevRel organization. So, I joined Google almost four years ago. And, you know, they're like this
thing called DevRel. And when people describe some aspects
of DevRel, I've been doing that my whole career in one way or another as either an
engineer, engineering manager, engineering leader at a company. All aspects of bringing a product to
market involve some of the things that I see people in DevRel do. So it was a very natural fit and honestly,
a really nice way to focus. My background is, maybe started in
customer service, answering telephones, helping people with their one server
or web host, like, "Your server's totally going to crash,
but it's going to be okay." And then becoming a system administrator,
learning to program, and then working in the open source world. So from Puppet configuration management,
through the whole container thing with Kubernetes, and now I'm at Google. - Excellent. And so, to kick it off, for those of you
who know Kelsey, I mean, he's well known as a fantastic speaker,
he's always on the keynote stage, and we'll hopefully get to in the end,
you're famous for your doing your live coding, taking big risks,
and having them really pay off and really making it exciting for the audience. So, as part of that, I want to start...
I think part of your strength, in my opinion, is that you're
such a communicator, you're such an educator, right? So, when you come out, you really make it
very clear that you have an agenda that you want to share, and you want to get
people to a particular place. So, can you share a little bit about that
passion, and how maybe you got those skills to be so good? - Yeah. So I think when I started
to give talks, I tried to mimic what a good speaker's supposed to be. There's like all these online tools and
books about how you should speak and how you should look, what you should wear,
and what you should do. And that never really worked for me,
I just use my vernacular, like, use my slang, there's some words I
can't pronounce and I'm okay with that. So, what I've resorted to doing is
just learning in public, okay? There has been some keynotes that I've
given where maybe I'm in a ChefConf and Chef is about to
announce a new product. And I get wind of this new product,
it's like, "That sounds way more interesting than the keynote
I have planned. Can you teach me what this new thing is,
and then maybe I can get that Hello, World live demo?" And it's one of those things where you're
in the hotel room, and you figure out how to get it to work, and you
get that happy moment. All I do is try to remember that and then
go and recreate that on the keynote stage. So for a lot of times, it feels authentic,
it feels like it's very clear because I don't stop working on that until it
becomes clear to me. And then I don't forget what it's like to
be a beginner and just take people on the exact same journey that I went on,
but you just do in front of them. So, I'm just learning in public,
and everyone has a different way of doing that. Let your personality shine through. And I think that what makes people
remember you because it's you giving the talk, and not the thing that
someone wanted you to be. - Yeah, so as part of that,
as I mentioned, the theme is "Knowledge Diversity." So that's such a great thing,
knowing that you're kind of leading people and I'm on this same paths with you. And in our space, as I mentioned,
a lot of people have to be quite knowledgeable and skilled to even
understand the basics of Kubernetes and know what you're about to talk about. So, do you encounter a wide range
of people with a wide range of knowledge, whether they're in the customer end,
because you do meet with customers, or as a product person,
or in the community? How have you worked
each of those engagements? - Yeah, so in my career at Google,
I work with a lot of developers and the community operations people. There's one aspect of the community that I
like the most, that's the people in the community that are willing to pay you
for your product, we call them customers. I like those community members too. And even in that context,
you also have their executives, right? So, this would be
the Fortune 500, the C-suite. They don't care about Kubernetes,
they don't care about containers at all. So, in their world, you know,
a lot of us build these demos for our products, they're equivalent
to the Hello, World. Anyone ever build those,
like that Hello, World, that initial wow moment? So the things I've been training myself
to do in that audience, right, it's a different perspective. But I come up with this idea
of Hello, Revenue, right? They want to see the thing that actually
changes the way they do business. And for a lot of businesses,
revenue is the top-line item. So, for me, is I make sure that I
understand the low-level skills, this is why I still contribute to
open-source projects, I take the time to keep up, keep my own skills sharp
by touching the keyboard. But before I go into an exec briefing,
I also read or listen to the last couple of earning calls. I need to understand where their
businesses so we can go from top to bottom, and it just keeps you prepared. So you got to stay sharp and you got to
make sure that you're paying attention to the environment that you're in. - So actually, as part of that, we talk
a lot about imposter syndrome here. In my early days, I was at VMware,
and we created these VMware, sort of, "cheat sheets" because we knew that
there are many CIOs who probably knew as much as they needed to know
about virtualization. But there's this whole emotional level
that they, like, devour these, you know, cheat sheets, because they still felt that
they didn't know enough. Is there something that you do in that
space in your way of addressing that when you go into those conversations? And in some cases, quite frankly,
people might be intimidated, like, "Oh, my God, it's Kelsey." I don't know. How do you create a relationship? - So, one, you got to get to the
place of confidence for yourself, right? Once you seek permission to dress
the way you want to dress... like when I go to an exec briefing at the
Fortune 500 and I'm not wearing a suit, that's just going to be what it's going
to be, I'm going to find comfort in this outfit. And even though I'm still uninvited,
I get to do Google keynotes on the grand stage the way I want to do the keynote. And over the years, you just get this
level of confidence that you have for yourself, everyone's going
to get in a different way. And when you're confident,
you tend to make people around you a little bit more confident. Because when you're confident,
you don't have to bring anyone else down. You're okay with other people being happy,
you're okay with other people being successful, you don't have to take
their success from them. So getting yourself confident first is
kind of a prerequisite in my book to make sure that you can
sit at that level. And then a lot of times when I'm coming
in, and, you know, people put me in this pedestal of, "He's the Kubernetes expert." And the thing is, I'm learning like
everyone else, some people just started in different points in time. Lots of people know a lot more
about other things than I do. So you just got to make sure that that's
clear, you can remind them of that. And then sometimes I'll also say,
"You may know this already. I'm just going to restate it
because it's necessary for me to frame my conversation." So a lot of times... And then you just be a human. Look at people when you talk to them,
give them body language. My goal is not to talk down at anyone. This is sometimes where I like to have
conversations with the whiteboard there. Maybe you ask them to draw. And then when you draw on the whiteboard,
then you're demonstrating that you're listening, so people understand
that you're not coming in to tell them what to do, you're just
coming to collaborate. And for humans, when that happens,
you tend to lower our guards, and then you start to have
the right conversation over time. - Excellent. So as part of that, do you have official
mentorship responsibilities that you do? Where do you extend that passion? - Yeah, so the mentorship thing,
like the official thing of mentorship to me is, I tend to deal well when the
people around me are happy and successful. So if you work with a group of people
who are down all the time, then you may not want to be around them. So if you have a team of people who don't
necessarily have that confidence, maybe they're not executing
at the best of their abilities, if you have ways that you can help at that
point in time, you have the ability to be a mentor. "Hey, I noticed that you're not leveraging
the best of your best skills. Here's what they look like from my view. I'm only one input. But if you did these things,
I think it would be amazing. And I'm willing to listen and collaborate
to help make you better." So, I think mentoring both ways. I also identify mentors. There's amazing people at Google who know
lots of things about different things. I take every opportunity
to come in and listen. The same thing about an executive. "Tell me about your business. What's challenging about your industry?" So you can have these little micro-mentors
that help you kind of level up. But I think if you want to grow,
you're going to have to help other people do the same, and I think you
got to make time for it. - So, as part of that, we talked
a little bit about sort of the empathy work that you put into your
customer experiences. Can you tell us a little bit about that? Because that's kind of a mentorship
through a program that you created. - Yeah, there was a lot of questions
I heard from some speakers like, "How do you get the engineers to buy
into the output of the DevRel work?" Right? When we go out in the field and we
come back with some of these learning and finding, the thing you hate the most
is that you don't want to pin these things to someone's backlog, right,
they already have a lot of work to do. Why would they re-prioritize this thing
that you're bringing them? Right? Every week, every day,
you're bringing these new things that are just being appended. How will they prioritize? And again, since I'm learning this whole
DevRel thing, I've been trying some new things. So naively, I go in and say, "Hey,
the best way, I think, for an engineer to understand the impact of this situation
is to make them be a customer." So, for the last four years at Google,
I've been working on a thing called Customer Empathy Sessions. And the very first empathy session...
so I can describe how they work, I was a contributor to Kubernetes
before joining Google. And when I joined Google,
we were trying to figure out, why is Kubernetes so hard? And lots of people
didn't agree it was hard. So I got all the core engineers of
Kubernetes, GKE is our commercial product, and I got them in a room and said,
"All you have to do is install Kubernetes. That's it. You're the core engineering team." - How much time? - "You invented this thing." And then what you do is you show them,
like, "I can do it in 10 minutes," because, you know,
I rehearse and I remember. "I can install it from scratch
in 10 minutes. Here you go, I just did it. You guys can have 45 minutes,
get into groups of fours, split up the work. This should be easy. I'll be right back." Then you come back and you
make them present. And it's like, "All right. I know this is too easy. I apologize. So let's how far did you get?" And no one even got to the point of
installing Docker, and I was like, "Oh, this is terrible. If you can't do it, how do you feel
that you can't do it? So what you're asking the customer
to do, you can't do. So, at this very moment,
you are the customer. I just asked you to do something and you
have none of the excuses that our customers have, so why
couldn't you do it?" Everything that we listed at that
point in time, either turned into some product feature, or more
empathy going forward. So don't go on stage and say,
"Kubernetes is easy to install." We shipped a tool called Kubeadm. Two weeks later, the core engineering team
somehow found time to build a tool called Kubeadm that you run,
and it gives you a Kubernetes cluster. And I wrote the guide because I didn't
think it was just a product problem, there's an education problem. So I wrote a guide called <i>Kubernetes The
Hard Way</i>, that show people every step of the way so that they can get confident
on how to actually build a Kubernetes cluster from scratch. And that's the output of just
one empathy session. I've done dozens of them at Google,
and they have permanent impact on the individuals that attend, because now
they understand what it's like to be in the shoes of our customers. - Absolutely. There's so many directions we can go
from there, because I'm sure we're all doing that. And even a simple thing I felt in any kind
of instruction or anytime I ask a question, when I get the response,
the first sentence is, you know, "Oh, it's easy. Oh, it's really easy." And then, you know, what follows is three
sentences of something like, "Well, I didn't get it in the first place and
that's why I asked the question." So then when you come back and
your first sentence is, "Well, it's really easy," you're setting
up a distance, right? - Easy for who? - Yeah, right, exactly. So, part of that, I was just curious,
you know, how do you programmatize something like that? Have you turned it into this program, or
it was a sliver of a single session that then now someone else has taken over, or
are you personally engaged in these empathy sessions? - So I am super lucky. So this is the very first DevRel
org I've ever worked in, all right? And the person that I report to is my
partner at Google. We don't really do the typical manager
relationship, we're strategic strategists on what moves the needle forward
across multiple products. And when I tried out this idea and we
started to get a little bit of success with them, Greg Wilson...
he's actually here, he's probably hiring too, so...There you go. And Greg Wilson was like, "Hey, you know,
you can turn this into a program. This is how we scale this thing." So there's a woman at our job,
Kim Bannerman, she reports to Susan, who's also here. And Google's really good at looking
at some things that work and turning into something that's
a little bit more scalable. So the way we think about scaling it is
train-the-trainer programs, showing people how it works in practice,
letting them do it and giving them feedback, and then putting it on the
menu of services that you can get from DevRel. So then what you start to understand is
like, team off-sites, if you're an engineering director,
if you're a PM org, and you're looking to get solid feedback that lasts a
little bit longer than the friction log, these events are in high demand, right? And we can only scale it by having it...
you know, there's a form, format, how long they are, what we do in them,
and what the outcomes, and how we follow up. - So, we're going to be able
to open up for questions in a little bit. I just want to make sure Matthew
and his mic is available. Excellent. So, please start thinking
about your questions. So, around open source,
talk to all of us, because we're in that world. Have you had any issues with talking
about metrics within the org or how you split your time? You talked a little bit about,
you know, your metrics, but what can you share with the audience? - So I've been in open source
a lot of my career. There's some open-source libraries
that I haven't maintained. There's some open-source tools that people
use, where there's libraries. I've contributed to the core of Golang,
the Go programming language, Kubernetes. And from that angle, it is very much
a community thing. I'm scratching my own itch,
and I want to share the results with other people and it scales
that way, that's great. Then the business context. I typically am very honest
and pragmatic with myself. Would I pay someone to do something like
this, if it was my money, if I was cutting the check? It's very hard to get lost in the big
revenue bucket of enterprise and just feel that my role is necessary. But when you flip it around, you say,
"Would I pay for that?" So let's talk about open source. You come up with the most
amazing product in the world. And, in your mind, we're just going to
open-source and give the whole thing away. And you make no revenue. Now, I need to understand,
what are you doing? You say, "Well, I'm building these amazing
things that I'm giving away for free." So okay, so, you don't just open-source
something, you got to maintain it. There's community, there's stickers,
logos...someone has to pay for that stuff. So when it comes to open source, yes,
you can talk about the number of outside contributors, the velocity,
the releases, but sometimes there's that part of that community, again,
we call them customers, they're also part of your community. So one strategy that I've seen work well
is when you build out something that can be totally open source, it's okay to
make it work for paying customers first. It's okay to make sure that their use
cases are covered in a way that they're willing to pay for or cover before you go
out there in the community. It's also good to set direction, though.
in some of these projects. So, if you're an executive at one of these
companies looking at all the buckets that you pay for, and I come to you and say,
"This open-source project is attached to 15% of the company revenue." They may not care about the number of
contributors, they are happy with that particular metric because they're
running the business. But then you make sure that the health of
the project are healthy as well so all the other metrics contribute to that. But that revenue metric when
it's business is also important. And I see a lot of people consider it
something that should just be off limits when it comes to open source. Right? So that's been one thing that
worked well for me and it avoids all of this kind of tension in the room where
people don't understand why we're doing this with this particular
open-source project. Every project is different and every
project requires revenue generation. Just be clear that you're just sharing. We're sharing this library,
we're not trying to monetize it. But this other thing,
there are monetization opportunities and here's our strategy
to do so ourselves. - So, part of that, a friend of mine,
we were talking about an experience, which I'm sure so many people in this room
have had maybe recently or many times, where they're
trying something new, they got into a chat room,
they were playing with open source, and their first contact was
highly negative, right? Someone sort of talked down to them,
"Why would you want to do this? Seems dumb," or, "You don't have
enough knowledge," etc. What guidance would you give for that,
especially within the context of empathy, and, you know, you want to grow your
community and yet this happens so often? - I've been on both sides of that. So, as a person at Puppet Labs,
we used do a lot of our support right there in IRC. People ask questions. Eight hours of doing this, you go home,
maybe you jump on to answer a few more questions. Sometimes you're a bit tired, too. Sometimes you had a bad day,
you just had the performance review, you didn't get the rating you want. If someone asked about how to install
Puppet one more time... On the other side of it as a new user
going in, you're intimidated, of course, it's new to you. Again, this is why I think confidence is
key, be smart, and something else, and carry that confidence
into something else. You know that you're going to be new
and something going forward, but then have a little bit of empathy
on the other side of that keyboard. You don't know what state of mind
this person's in. Sometimes, you may be in a view that you
read the sentence in a way that you think you're being attacked and sometimes you
can ask clarifying things. And then the reality is,
there are some assholes in the world, period. If you happen to get one of these
assholes, you can definitely not continue to dialog with them,
but that's actually reality. So I'm never shocked
by this particular situation. And I think the last thing I do as a
person asking questions, I make sure I do everything within my
power to find my own answer. And before I lead in with the question,
I always say, "Hey, these are the 7,000 things that I tried and I'm still stuck. Here's my question." And some say, "Oh, I can appreciate that. You're linking me to a Google Doc,
you're showing me your experience, you've captured the errors. This is an excellent bug report,
by the way. Thank you for doing all of that. Sure, I can help you." So you're going to make sure, are you
bringing the same effort to the table and the ass that you respect
on the response? And if the answer is not there,
then a lazy question may get a lazy response. So I think it kind of goes both ways. - And, you know, to give some
benefit of the doubt, you know, I do know people who are good
people, and they just give a quick response on slack. And they don't hear the tone that maybe
actually went across, or they don't realize that a tone could be
interpreted in the words that they chose. - Don't be afraid to apologize. I've apologized publicly on Twitter
before where I've said something... I mean, I think I was giving a talk one
day, and I said a joke about PHP, the programming language, right,
the software, not anybody that works on it. And I got home and there was this email
about how that made someone feel. I was like, "PHP?" And the email was about, like, "Hey,
you know, some of us aren't so fortunate that we get to work in other languages. I'm stuck with this PHP,
and I just don't feel good anymore." And I was about to do this quick response,
like, "Dude, you have to go outside and do something,"
but then I thought about it. The context that this person
was in, that wasn't healthy. So I said, "You know what,
I apologize, and that's good feedback." And sometimes you just got to leave it at
that, even though I want to educate this person about detaching their identity from
their work, because the work will change and you will remain, and you
shouldn't be that easy to prod, but that was inappropriate at the time. The apology was the only thing necessary. So, that's kind of the way I do it. So, I'm okay to apologize if I need to. - One question. We have a hand over here. I don't know if there's a hand over there. Maybe two, we'll see. - [Audience Member] Hey, Kelsey.
My name is Suyash. My colleagues went to QCon
and they were blown away by your talk. They were like, "Oh, this guy..." And this lady, she's a pure marketing
person, she's not a technical person. But she was like, "This guy
gave such a great talk." What is the secret to doing talks,
the ones you do where a sold-out crowd and people are leaving
with this kind of feeling? - So I'll tell you. So one thing is I stopped using
slides a while ago. Number one, I'm not good
at creating slides. There's nothing wrong with slides,
I'm just not good at creating them. And then also, I don't write the talks
ahead of time and I don't practice the words ahead of time. Because I found some confidence
in creating these demos, they're actually micro-demos. I have hundreds of them and I play them
like a musical instrument. So when I'm on stage and I'm talking about
a particular topic, I can swap in and out of them as I see fit. And the way I see fit is by looking at
people and their body language. I remember when I was talking about
Kubernetes one time at an event, and I don't think people understood
what the Scheduler was. And I had Tetris and a Nintendo emulator,
and I played a game of Tetris to explain bin packing to them. And when the body language improved,
I knew I had people at the right baseline to continue on. So then I put that one in my back pocket. So typically, when I'm given the keynote,
you scan everyone in the room. I assume everyone is here for a reason. I don't care if you're an engineer or
are in marketing, but I need you to leave with something. Either you're going to be inspired,
maybe you get it just a little bit more, and I'm willing to adjust that talk
mid-flight until the body language improves. So when I see this person where that light
bulb goes off, we're going to double down on that. And that's how I want you to leave. So, I've seen this work a number of times,
I gave a keynote, I thought I was doing terrible. Three minutes in, someone I knew
just got up and left. I was like, "Come on." This intro was crazy. I just said, "Okay, Google,
critic cluster," and it came up and you walk out? No, hell way. And luckily, I checked my Twitter feed
and that person said on Twitter, in the first four minutes, he was inspired
to go build something new. And he went and built it. So I'm like, "All right, we back." I think we have one more here. Maybe we just hand the mic to him
and then we'll wrap. - Just before we go into that,
though, we talked about jazz, right? Because I said we wanted to make sure we
ended with some bits about how you're such a master of live demo and how,
you know, you come into this. And I felt one of my observations was that
I saw you do a demo that wasn't going well. And I've seen others in the same situation
where there's a wall and they're sweating on the other side of the wall going,
"Oh my god, I'm supposed to be the speaker. I'm supposed to be prepared. I'm supposed to be, you know,
the one with knowledge. And this isn't going well." Whereas with you, you're like,
"Cheer me on," right? You create a team, you're like,
"If this works, we're all going to celebrate," right? - I'll tell you. The very first live demo of Kubernetes,
we had it somewhere in San Francisco, the entire Kubernetes team here, Red Hat,
CoreOS, all these colleagues are here. They're like, "Kelsey, we want you
to do the opening theme because people don't know what Kubernetes
is and you're going to show them." I was like, "Oh, hell yeah." I pulled my laptop. I'm using VMware Fusion. And it turns out in VMware Fusion when you
disconnect the network and you connect to another network,
all your VMs go to hell. So, I'm like with all my swag, like, "Ah." Then you just do, "Ah, ah,
ah," and they were like, "Two minutes." And everyone was like, "Oh, this is done. He failed." I was like, "Time out. Does anyone want to see this work?" And they were like, "Yeah." And they started leaning forward. And then I said, "All right,
we're going to delete all the VMs. It's Friday. It's an outage, and the customer's on the
phone, you're on the bridge." And like, "How's it going?" He's like, "It's cool." "And you delete all the VMs and we
bring it up step by step." And I remember the VM launched
and people started clapping. I was like, "You guys are clapping
because I lost the VM? This is going to be good." And then the Kubelet comes up,
Docker comes up, and you're like, "This ain't going to work, but who cares?" Kubectl apply, the thing runs,
and everyone cheers. And then it taught me that people are
actually on your side. If you look at them,
and you tell them where you're going, more than likely, they'll
help you get there. - Last question. - [Audience Member] Test, test, test. Hi, Kelsey. Daniel [inaudible] here. Just a quick question on the theme
of knowledge, on diversity of knowledge and whatnot, and confidence. What was your learning journey in the
realm of...in the context of confidence building? What were some of the early, sort of,
moments like this anecdote that you just shared? What were some confidence-building
anecdotes along your journey that helped get to where you are today? - I'm getting emotional now. I think I was about nine,
and a single mom. And, you know, you got a single mom,
it's like, "Where's my dad?" "I don't know." And then single moms do everything,
but you don't know how much at nine. And I remember my mom was working at the
dining room table, she bought a little typewriter. And she bought, like, how to type,
it's like a flip book, right? You learn the touch type. And my mom is sitting there,
she used to be a seamstress like in a factory. You go in and you sew clothes all day. And I remember this because she used to
sew clothes on the side and sell them to the daycare. She would come and,
"Who wants to buy things?" How you make ends meet. So that's the job my mom had. I can tell that wasn't a high-paying job
based on the environment you live in. And you watch your mom learn how to type. There's no person you ask. And remember, it costs a lot for a
typewriter when you're trying to figure out do you eat or do you type? And she goes through and she learns to
type and she interviewed for a job with the US Customs. And on the resume is like,
"How fast do you type?" And she was like, "Ah, you know,
what's a good speed?" She was like, "Yeah, that plus five." And they were like, "Okay,"
and she was like, "Yeah. So we need to figure out how to type." So she's hunting and pecking. She's hunting and pecking. And then, remember, you're a kid,
and you can hear the flow change, you hear, "Ti, ti, ti, ti," then you hear,
"Ti, ti, ti, ti, ti, ti, ti, ti..." Someone's getting good at this. And then, remember, your timeline
as a kid, you just see your mom come home with a different uniform on,
with a different look. Now she has the badge, nine to five,
not nights and weekends. And you understand the power of learning,
the power of, like, learning how to do something, and you also learn who's
responsible for learning. So, when I look at that situation,
and when I'm on stage, I'm always looking through
the YouTube video. There's someone that I've met that someone
who's like, "Hey, man, don't tell anyone. I just got out of jail." I was like, "They let you
watch YouTube in jail?" And this dude was watching YouTube videos. And he came out in the job market. Little did he know that Kubernetes is hot. And he just went for one of these jobs
that said, "Kubernetes experience," he's like, "I got that." It's like, "We don't believe you
because no one has it." And he was able to demonstrate
that he has Kubernetes. And that startup got acquired. And he was engineer number one. And his boss was like, "Thank you. I don't know how you
met him or taught him, but now he was able to change his life." That to me is the power what we do
and that's what keeps me going. - All right. Can't find a better way to end that one. Thank you so much, Kelsey. ♪ [music] ♪
