Facilitating the Spread of Knowledge and Innovation in Professional Software Development

Write for InfoQ


Choose your language

InfoQ Homepage Presentations Passion, Grace, & Fire - The Elements of High Performance

Passion, Grace, & Fire - The Elements of High Performance



Josh Evans talks about the elements of high performance: passion, grace, fire, and what matters when trying to build and shape teams for high performance. He focuses on a period at Netflix when demands on engineering to quickly deliver multiple, parallel, large-scale technical transformations was the norm. The transformations enabled a global, scalable, reliable, successful streaming platform.


Josh Evans is a 17-year Netflix veteran (30 years in tech) with leadership experience spanning e-commerce, video streaming, infrastructure, developer tools, and technical/cloud operations. After Netflix, he spent a brief time at GitHub, establishing a new data platform, and is currently advising a video streaming analytics/automation startup called Datazoom.

About the conference

Software is changing the world. QCon empowers software development by facilitating the spread of knowledge and innovation in the developer community. A practitioner-driven conference, QCon is designed for technical team leads, architects, engineering directors, and project managers who influence innovation in their teams.

[Note: please be advised that this transcript contains strong language]


Evans: I was never going to work in technology. Never. Growing up in New York my father was a mainframe programmer, he worked for banks and insurance companies. He worked in an environment that didn't look exactly like this, but pretty close. Those large printers with the green and white paper, the very large mainframe computers. He worked in COBOL, CICS, those kinds of technologies, and because of his bias, because of the way he felt about his work, it really affected me.

When I went to school in 1983, I went to UCSC at the ripe old age of 17. The last thing I wanted to do was work in technology. I had friends who were doing computer science degrees and I was, "Good luck with that. Not interested at all." I wanted to do music. Even though I didn't really have a strong background in music, I wanted to do that.

Then, I wandered off and I started focusing on politics, political science, social sciences. Eventually, I settled on art, and something a little bit more arcane that you would expect, printmaking specifically, something that I had done as a child and really enjoyed and went back to. That ultimately ended up being the thing that I focused on because I really valued the arts and all that.

Something happened in 1988. I was on the seven-year plan at UCSC which means I, sort of dropped out, couldn't go full-time, needed to support myself and went back part-time. I was waiting tables, I was cleaning houses, and really just dreading that work. I had the good fortune to meet a gentleman named John Seamster who was the Director of Sales & Customer Service at this company called Borland. The time they were Borland International they were writing the personal computer wave and they were trying to go head to head with Microsoft. You can probably tell how that story played out based on the fact that probably many of you here have never heard of Borland, or if you have, it's just some word.

It was very cool to me going in there being exposed to personal computers, being exposed to a more modern technology. Before I knew it, I was done with my art degree and realized that was an extremely impractical decision and needed to do something that was actually going to make some money. I moved in across the hall from customer service into tech support. I'd started out a very low-level job but I'd learned just enough that they were willing to give a shot to somebody to try out a tech support role.

Within a short period of time, I had really gotten pretty engaged. I had moved into doing tech support for a product called Paradox for DOS. Paradox was a LAN-based database. This is before the internet so your only network was on-premises and if you're lucky, you could have some kind of WAN that would connect you to other locations.

The support that we did on CompuServe was fairly straightforward but the tooling was terrible. CompuServe, for those of you that might have seen old school bulletin boards was all text-based, it was hard to search, it was hard to collaborate, there was poor threading support for discussions. It's hard to know if somebody else had already answered a question that you thought you might need to answer, and so it was a pretty clunky experience. I decided I was going to try to fix the problem.

I had never written a line of code, but I decided I was going to try. I started reading books and I ended up building an application I called Quicksilver and it was essentially a database messaging tool that would shell out to DOS, it would go and download messages, parse them, integrate them into a database, and then make them available for collaboration. It was a lot of fun. I was spending nights and weekends working very hard because I wanted to, because I was really engaged.

I guess the moral of the story here is never say never. You don't really know what's going to happen. Also, something changed for me during that time period that took me from one perspective to another and helped me find a passion. This is a passion that I've now had for 30 years.

My history here: I was at Borland first for about eight years. I moved through a bunch of different roles there from tech support. I wrote a test automation harness for Paradox to Windows, wrote a bug tracking system, dabbled in management. I went to a failed startup after that which I won't even mention. Nobody here would have heard of it anyway. Then, I went to this place called Netflix in 1999 and I think really for me that was the beginning of my career in my mind.

It was an amazing experience. I started in eCommerce, did that for six years as an engineer, moved into management, and then after that, moved into managing playback services for streaming. Then, after Netflix, I spent some time at GitHub and did some work there, but now I'm in a phase where I'm doing more advising, consulting, and those kinds of things. I'm spinning up an effort to some consulting firm called Technical Leadership Strategies to help others grow their careers and grow as individuals.

Passion, Grace, & Fire. This came from a song written by the gentleman on the left here. His name is Al Di Meola. Al Di Meola wrote this song I want to say 30 years ago, probably longer than that. He's in his 70s now and he probably wrote this when he was in his 20s, probably about 40, 50 years ago. Beautiful song but also the title of an album done by these individuals here, three virtuosos. They were called Los Tres Guitarristas. This was in Barcelona. They did a world tour together. Each one of these individuals was a virtuoso musician. Extremely fast, extremely good at what they did, and they were innovators in the jazz fusion space.

Al Di Meola brought his own special flare: Mediterranean, flamenco, rock, jazz. He went to the Berklee School of Music and he had some serious chops. In fact, at 19 years old he joined a band called Return to Forever which was a prototypical jazz fusion band, very formative. Then, he went onto do great solo work with many great musicians over time.

John McLaughlan, also a jazz fusion specialist was more focused on East Indian themes, classical music and flamenco, so a real variety of different sources. He worked with Miles Davis early on in his electric albums, and he formed a couple of different bands over the years that were focused on East Indian themes mixed with rock, blues, jazz, all of that. One called Mahavishnu Orchestra which I listened to in the '70s, believe it or not.

Then, the third musician was Paco de Lucia and he was from Spain and he was a flamenco guitarist who started merging and innovating by integrating classical themes and jazz into his work. He was famous for picados and rasgueados. This is the strumming flamenco style mixed with jazz leads and go back and forth between the two.

The Elements of High Performance

When I think about this concept of passion, grace, and fire, when I think about these virtuosos coming together doing the best work of their careers, they certainly did a lot of interesting work afterwards, but this was sort of a high point for these musicians. They only did two albums but they're pretty amazing. We were listening to some of their music when all of you were sitting here waiting for the talk to start.

The way that I map this concept of passion, grace, and fire to what's relevant to us here in the room is by looking at these elements and breaking them down and thinking about the things that we care about in the work environment when we want to inspire high performance.

Passion really speaks to your motivations, your affinities, your drive, and early formative experiences that give you the motivation to do great work. Grace is about the skills and the capabilities, the experience, and also how you carry yourself, who are you as a person, what kind of integrity do you have, and how do you interact with other people, which can also affect the way that you function. Then, fire is about bringing it. It's really about tenacity and delivery. A lot of people have great skills but they also need to bring things into production and have that done well. I think that this premise here, this is the feeling that we want to have when we have high-performance teams, this idea of passion, grace, and fire for me is almost an archetypal kind of meme.

Then, the question is, now that we've had all this set up, how do we leverage this concept of passion, grace, and fire to inspire high performance and actually make that impact? As leaders, that's what we're looking for.

That's what we're going to talk about today. I'm going to start by talking about the environment in which we operate because culture and company is critical to the success if you want to try to create a company of owners. Then, I'm going to spend some time talking about some of the best work that I feel like I've done, the best experience I feel that I've ever had working with a team of amazing people, the period of 2009 to 2013 when I managed the playback services team. Then, I'm going to use that just to flesh out and give you more of those examples. Then we'll talk a bit about your journey, how you can take what I've talked about and apply that yourself.

Ownership-Oriented Culture

Ownership-oriented culture. Seeks to foster a strong sense of ownership in every employee, have every single person feel like they are an owner of the company. We can do that by igniting passion about the work. Fostering grace, giving people the ability to develop new experiences and new skills. Demanding fire, holding people accountable and making sure that things get delivered. The big question is, how do we do that?

I'm going to do a quick detour. I'm going to talk about a book written by a gentleman called Daniel Pink. I think it was about 15 years ago. He talks about the surprising things that motivate us. One of the fundamental questions of his book is how do you motivate people, especially in the work environment? Does more money actually lead to better performance?

Here's what's interesting. There's a study done by MIT that Daniel Pink refers to and they break it down into two different dimensions. I'll test your aptitude here. I'll test your awareness here about what this is. There's two different types of work. The first kind is manual, repetitive, algorithmic work. This is work that's already been figured out. Most people come in and they just do their process and they know what they're doing every day. There isn't a ton of creativity involved.

How many people think that paying people doing this work more money will actually give you better work? Ok, so you do get incremental improvement. You can actually get people to do better work with this kind of work. The opposite is true for the kind of work that all of you do. If it's cognitive work, if it's creative, if it's problem-solving, if it's innovation you actually will demotivate people by trying to give them a bonus to do better work.

Going back to this, the fundamental premise of Daniel Pink's work is that drive, this passion, this thing that makes us do great work is fueled by three things. One is autonomy, the freedom to go and pursue a project and do it in the way that you think is best without being told what to do. Mastery, the ability to develop new skills which is highly addictive as all of you know when you start learning new things it's pretty exciting. Purpose, this is actually having something that needs to be accomplished, accomplishing something that matters. That's the impact that we're talking about.

A great example of this is open source. People do this work on nights, on weekends, in their free time not because they're getting paid to do it but because they want to do it, because they feel like the work matters and it's exciting because they can develop new skills.

For me, going back to my story, I fell into this kind of environment by accident. I joined some company without really knowing what it was about. I got into tech support and before I knew it I was learning how to write code, I was building tools, I had a supportive manager. I got totally hooked on learning how to program. I wrote my first parsing routine, I learned about recursion. It was a really exciting time for me. The purpose was very grounded. It was like, I'm solving a problem for myself and my coworkers, and what could be more satisfying that fixing a problem, making things better?

There's one other piece to this is and this is not what Daniel Pink talks about. I'm adding my own editorial here which is that formative experiences also matter. My love of print-making taught me a love of craft and a love of what I like to call the zone. That's when you're working for hours and you don't even realize that time is passing because you're so absorbed in what you're doing.

The other thing that was formative for me in this example was that around age 11 when we moved to New York I discovered science fiction and I was reading Isaac Asimov, Robert Heinlein, and Frank Herbert and I fell in love with this idea of the future and inventing the future. Despite all my biases going into the decision about whether or not I should get into technology, there was a fundamental part of myself that I tapped into that ultimately led me to the career that I have today.

We're talking about passion, grace, and fire. We're really talking about in Patty McCord's terms, and Patty [McCord] was the Chief Talent Officer at Netflix for 14 years and a mentor of mine, somebody I very much respect. She wrote a book recently about a year ago "Powerful" that touches on some of these themes. What she said was, what she came to understand in a new way once she got into the scrappier startup world is that people already have power coming in.

A company's job isn't to empower people. You should hear Patty [McCord] use the word empower with dripping sarcasm. This is one of her least favorite words. Because people have power coming in. The goal of the company is to remind people that they walk in with power and they create conditions for them to exercise it, not to empower them and give power to them. It's more about getting out of the way than it is about doing something proactive. If you do this, this is where the amazing work kicks in.

This concept of an ownership-oriented culture, we talked about this before. This is accomplished applying the Daniel Pink methodology by aligning passion with purpose, by providing opportunities to autonomously master new skills, and embracing accountability. At the end of the day, you do need to make sure people are bringing the fire.

Of course, this is all an indirect reference to what's obvious which is Netflix Culture, which is the culture of freedom and responsibility which is a great example of an ownership-oriented culture.

There are four concepts related to this, but first and foremost, if you go to Netflix and you talk about freedom and responsibility, the very first phrase that comes after that is context, not control. This is the management strategy. The context not control concept is supporting by four pillars. One is something I'll call extreme transparency, the next is deep delegation, radical honesty, there's a whole book on this, and true accountability. I'll break this down for you.

Extreme transparency. I'll give you a great example. In 1999 when I started at Netflix, they had a weekly meeting outside by the picnic tables where leaders of the company would come out and just talk about what was going on in the business, what was the total number of members, what were the challenges we had, what was in the news, etc. That helped everybody stay in sync and be aware of what's going on.

Netflix took it to a whole another level over the years. When they started doing quarterly business reviews, directors and above would meet off-site, they'd create hundreds of pages of documentation about what happened the last quarter, what's going to happen in the next quarter, what strategies are being applied, what worked, what didn't work.

Then, after that event was over, they would go back to the company and share virtually all of that documentation and they would go to a roadshow and do Q&A and talk about all the important things that were going on in the business so that every single employee knew pretty much as the directors and the VPs and the CEO. This is how you develop some of that sense of ownership.

Deep delegation is deep in two ways. It's deep in the sense that it's a complete handoff as much as possible to the person responsible for doing the work, not one of these, "I'll hand it off to you but I'll micromanage you a little bit," but much more of hands-free kind of operation. Deep in the sense that you want to push the work to the deepest level of the organization. You want experts doing the work. You don't want managers designing systems, you want engineers who deeply understand how to do that doing the work.

The delegation to managers, to give you two examples, was the ownership of their own team. Every manager's responsible for the success and the composition of their own team, and this means hiring, onboarding, coaching, setting vision, defining strategies for the team, putting roadmaps together, all of the things you would think about wholly owned by the manager. They might get some support to do those things, and they frequently do, but they're responsible for it. This is a clean handoff.

The other is each manager, in order for them to be able to do their work, they need to be able to delegate completely out to the people on their team so that they can scale their team.

Great example of this kind of thing - complete delegation, "You Built It You Run It." Netflix adopted this because it was such a natural thing. If you own a service completely and totally you're going to do a better job than if you share that responsibility with other teams, it's unclear.

Radical honesty, the third component here, the third pillar is really just about giving honest feedback and doing it in a way that people can hear. You want to be blunt and direct but compassionate. You want to find that right way to give that feedback, not so you can feel good about giving the feedback, but so you can actually make an impact on that person and help them grow. You're acting as a mirror.

Then, finally, true accountability. This comes in a couple of different forms. Initially, of course, you want to give people feedback. You want to make sure that they know that you know that things didn't go well and talk about how you can do better. That's always a good place to start, but at the end of the day there is this saying in the Netflix culture deck, "Average performance results in a generous severance package." Netflix is expecting people to really bring it, to really be high-performance and average just doesn't meet that bar.

Of course, the opposite is true that people who really excel in this environment will be rewarded with more responsibility whether that's a management promotion, the ability to run larger projects, do cross-functional leadership, and that creates a set of incentives that really reinforce the culture that you want.

The last piece is something Patty McCord always talked about, mature fully formed adults. How many people work with other people that are not mature fully formed adults? There's a lot of them out there and I'll tell you, age has very little to do with it. I've met people in their mid-20s who were amazing and mature and then there's people who are in their 60s and 70s who are not. It is a character thing, it's not an age thing, from my perspective. Fully formed adults are humble and introspective. They're hungry for the feedback, they want to do better, they want that mirror and they want to grow from their failures. They want to learn.

They have integrity and they understand that when you work for a company, you're a steward of that business. Yes, you have to think about your own career, you have to think about your team, but at the end of the day, it's about making the business successful. You need to brave and direct and honest with your peers. The last piece of judgment. You want to right most of the time. You don't want to have somebody who's throwing spaghetti against the wall but somebody who's really thinking about, "Here's a problem. Here's how I think we should solve it. Here's the data. Here's my rationale. I have good judgment about that." These are the kind of people that can be successful in an ownership-oriented culture.

A Team on Fire

Let me give you an example of the team that I've worked with from 2009 to 2013. Let me give you a little bit of context around that before I jump in. Some of the challenges that I encountered when I took this team on early on. In the fall of 2009, when I took over this team, they were actually called Electronic Delivery at the time. That was because we were trying to differentiate it from DVD By Mail.

I renamed the team and it went through some evolutions which you can think of as playback services. The devices supported at the time were Windows Media Player which was transitioning to Silverlight at the time. We were also on the Xbox 360, the first gaming console that Netflix was able to get onto. Of course, Microsoft had leverage so they had an exclusive contract around gaming consoles which I'll refer to in a little bit. There was the ROKU Player which was the first CE device which could act as a Netflix player, and then Netflix was starting to aggressively move into the more turnkey consumer electronics space where they could try to rapidly get Netflix onto the digital televisions, set-top boxes, at the time, Blue-ray players, etc.

When I came in in 2009 there were two projects underway. The team was small. There was only about five or six people on the team and they were delivering two things at the time when I joined. One was the disc-based version for the PS3. Because of that exclusive contract with Microsoft, in the second year of that contract, we could only do a disc-based solution and then after that, we could go downloadable. That was the disc-based, had to be physically shipped to customers and it was like a game and you plug it in and install just as if it was a game.

Then, we did this rather painful project called BIVL. Sony BRAVIA systems had this very proprietary user interface where you had to shove stuff into their environment. It was a real bespoke application, very challenging, but did get our foot in the door with Sony and helped us forge a good relationship with them.

I was working on that. We had engineers with in progress, we had launch dates. I thought I was getting up to speed. A couple of weeks into this, Anthony Park who's a director, I believe, at the time who comes to me and says, "We're thinking about international and right now, all we do is English. We want to do subtitle and alternative audio and those kinds of things." I said, "Cool. I think I've got an engineer who can work on this," and I committed to it prematurely.

Next day, a gentlemen named John Funge comes to me and he says, "You realize as soon as you're done with all of this, we need to do the disc-based version for the Wii." It was a project called Link at the time. "It's going to be all hands on deck and it's going to be an aggressive schedule," and all of that. "Oh, shit. Ok." I took a little bit of a step back and I came to realize what I had just stepped in being moved over from eCommerce and moving over the streaming.

I came to realize this wasn't going to stop, and I was only thinking about one track of many that I needed to be able to do over the next couple of years for Netflix to accomplish what it needed to accomplish. They wanted to be on every possible device in the world. There were ongoing sets of features and security requirements for contracts with movie studios. We wanted to be a global company so we needed to start working on internationalization.

There's this little cloud migration thing. I think some of you might have done this before, it takes a little bit of time to work through that when you've got a monolith and you're trying to turn it into a whole new architecture in the cloud. Then, of course, we needed to learn how to run at scale and we needed to learn how to run at scale in the cloud which is an effort unto itself.

My reaction: I had a total meltdown. Literally, after I had the conversation with John, within an hour I had this raging headache and I was like, "Oh my God, what did I get myself into?" I went home. I'm literally lying on the floor and my wife is talking me down. I'm, "Ok, I don't know what I'm going to do. I think I made a big mistake taking this job."

The next day, I went and sat down with my boss, the person who had decided to make this change and asked me to go take this team on. I said, "I'm panicking. I don't know what to do here. I'm totally overwhelmed. I think you made a bad choice here. I don't think I'm your guy." It was funny because the smile on his face was exactly how he responded to me. He looked at me and he said, "Josh [Evans], it's because you panicked that you're the right person for the job. It's because you care." I was, "Ok, let me run with that for a while and we'll see how that goes."

The first thing I did was I found a willing victim to delegate to because I wasn't scaling and it was pretty overwhelming at the time. We had a lot going on. I needed to start figuring this out, so I started practicing deep delegation. I picked somebody extremely capable. I got very lucky with this team that I took on, some amazing people already there.

Philip [Fisher-Ogden] had come from Lawrence Livermore Labs, a goldmine actually for us at the time. We found five or six just absolutely stellar people from there. Strong computer science background, had been a teacher for one year teaching math and science. He had a year of streaming experience which at the time was head and shoulders above pretty much anybody out there because there weren't a lot of companies doing this at the time. He already had written a mock version of our primary server for testability. He had a strong technical operations focus. He was the guy that got called when nobody else could figure out what was going on, and he was methodical and systematic.

I leaned on him pretty heavily, and for the Wii project I made him the team lead and he nailed it. Then, I decided at that point to promote him to be a manger. Right around that time Apple came to Netflix and said, "We've got this new device we're going to launch. It's called the iPad and we can't imagine shipping this without Netflix pre-installed on the device." They saw Netflix as the killer app for the iPad when it first launched. Philip [Fisher-Ogden] took that on, and not only did he take it on but he single-handedly did all the project management, he figured out the technical specifications to put something on the iOS player, he trained and helped teams on the device side that were doing it.

When the demo wasn't going well the day before Steve Jobs was going to do the launch, he was the one on the phone troubleshooting it and fixing all the problems. That's ownership. That's real ownership. Then, he went on and addressed the other launches throughout the rest of the year. For me, I got to completely sort of relax on this one front and I got to think about other things for a while and that freed me up from a scale perspective.

It only started with Philip [Fisher-Ogden]. I realized that after I did that, that I needed a whole lot more than one person on the team, that we probably needed to double, triple, or quadruple the size of the team that we had, and there was no way I was going to be able to do this if I went and spent six months hiring managers or more. Let them start hiring people. I went with a promote from within strategy, creating a number of opportunities for master for the people on my team.

The next person that got promoted was Julie Pitt, a very smart brilliant engineer who had a real attention to detail and a passion for refactoring and code quality, and had a reputation for always leading the code better than when she found it. She took on a team that we called streaming server which was all those core functionality for internationalization. Think alternate audio and subtitles as an example.

Next, we had Ranjit Mavinkurve who run the team before me but then he'd gone and done a pilot project with the infrastructure team to move the first streaming device type over which is was Silverlight from data center to cloud and then he came back to the team to run that effort for two years basically getting us completely moved over within that time period.

Then, the last star, last but not least - actually, there were many but I'm only focusing on these four - was Greg Orzell who had worked on a variety of roles at Netflix from IT to cloud architect. He was one of the original authors of Chaos Monkey.

Now that you've got leaders in place, the next question is how do you make sure every new person that you bring in is also going to have that passion, grace, and fire that's appropriate for the team? How do you make sure you're building a team of stars and you're really going to inspire that high-performance?

There is this thing that Patty [McCord] calls The Algorithm, and don't worry, there's no algorithm here. She liked to call it that because she thought that it would excite engineers to hear it phrased that way. Forgive the grammar, this is classic Patty [McCord]. I could never remember the wording. Thank God she wrote a book because I could go copy it, because I could never remember exactly how she phrased it, but the meme was always there in my head. "Is what this person loves to do, that they are extraordinarily good at doing, something we need someone to be great at?"

Mapping it to our passion, grace, and fire set of memes here. This breaks down very cleanly. It sounds all convoluted but at the end of the day, it's about these elements. What does this person love to do? Why do they like it? Do we need somebody to do this? The grace piece is, what is this person really good at? What skills do they have? What experience do they have? What are they bringing to the table? Is that something that we need? Also, what characteristics are necessary for success? You can think about those more soft skills, those cultural elements that are necessary for functioning well within a company.

Then, for fire, it's what's the track record? Can this person deliver? There are many people out there that I have met, especially more in the research space who have the skills but struggle with getting to get it out into production, make sure it's running, and make sure it works, kind of piece.

We crafted interview questions that allowed us to get at this, and I think that this is some of the meat that will hopefully be most helpful here. How do you get at this, how do you know if somebody's going to have passion, grace, and fire, and whether or not it's actually going to work for you in your team, in your company, and in your culture? A lot of it has to do just with simple alignment.

I like to ask people about what their origin story is, what got them into technology, sort of similar to my story so that I can try to understand what makes them tick. That's just good context to have. Then, I like to dig and talk about what's a project that you are really proud of. You can learn so much about somebody by asking them what they think matters. Can they ground in what was important to the business? Can they talk about impact or they talk about what they learned? All of these things are things that are worth looking into.

Then, of course, assuming they're here interviewing and wanting to interview at Netflix, you assume that they're excited so it's a good way to phrase it which is, "Why are you excited about this role?" Even more importantly is most people are interviewing in multiple places and the really telling question is if somebody's interviewing at Netflix and Google and who knows, Airbnb, and they're working in completely different spaces, how do you make the decision? This is where you can really get to what makes people tick. How are they going to decide which role they're doing? If it's a choice between eCommerce and streaming, does the domain matter or is it about scale and infrastructure? What is it? These are great ways of uncovering peoples underlying passion.

In terms of grace, I'm going to break this down into two elements. There's the technical skills which really matter and you do need to make sure you're asking about the things that are going to help people be successful. For us, it was about distributed systems at scale. We did have some tools folks and other folks we added over time, but that first year was who can deal with distributed systems and scale. I was really excited about working on that. We just focused on the fundamentals at scale: Data structures, algorithms, concurrency, logging, synchronization, bottlenecks, all of those kinds of things. Can people think about a reason through those challenges?

Then, we would have them design a system that solved the problem that we already had and to see how much they could flesh out the whole system. Do they ask good questions? Do they probe and push back on requirements? Do they think about operations, testing, instrumentation, all the things that we know need to happen? Do they think holistically and do they own the whole project or do they just take what's given? That would give you a good sense as well as whether or not they're going to be successful in an ownership culture.

The grace piece is all about whether or not this person will thrive in an environment that can be very challenging especially if you're not a mature fully developed adult. Describe a situation where you were in serious conflict with a co-worker, and what did you do? By the way, everything I do is situational and so everything is, "Here's the question. What happened? What did you do for the most part?" In this case, if you're in conflict, people will do one of two things. They might do a call to power and go to the boss and go, "I don't know how to deal with this person." They might also go and sit down with that person and say, "We've got a problem. How are we going to work it out?" The latter is what you're looking for. That's what autonomy looks like when somebody is a fully formed adult. They'll go and try to solve the problem themselves before they go and ask somebody else to try and fix it for them.

What's the most difficult feedback you've ever received? How did you respond and how did you apply it? By the way, "My greatest fault is that I work too hard," is the BS excuse, I'll stop listening at that point. Some people will be very candid and very genuine, and that's what I look for because those kinds of people can take feedback and do something with it. Finally, what was the most difficult feedback you've ever given? This is the opposite. Can you be brave enough to do it? This is a hard thing to do sometimes, to tell somebody bad news and find a way to say it to them that they can hear.

Then, describe your most painful personal fear. This gets to the deep introspection as well. Can they do that themselves? If I fail, does somebody need to tell me that I failed? Do I know? Do I go think about it? What happened? What can I do differently next time? Those are the kinds of people you want in this kind of environment.

Finally, you want the track record of delivering great work on time. This is the fire part. This is ultimately what really matters. This is why the business hires employees in the first place to get these things done. I'll ask them to describe a project that you were responsible for that was failing and then you turned it around. This gets at tenacity. Can somebody dig in and figure it out, recognize that it's not going well, catch it at the right moment? Get in there and figure it out. Instead of just throwing up your hands or doing a call to authority, giving up.

Describe a project that you owned end to end. This gets back to the design a system and think holistically. What were those stages? Can they think through the conception, the design, the requirements analysis, instrumentation, the delivery mechanism, getting it into production, instrumentation production, all of those things?

We did this. We put a methodology together, we held that high bar on technical skills. I think we did a reasonable job bringing in free and responsible people. I'm going to talk through just a few of the people that we brought in, give you some examples of what this looks like.

One of the first people I hired, in fact, I think it was the first person I hired in 2009 right at the end, was a gentleman named Ishaan Shastri. Ishaan [Shastri] was at Tibco and VMWare, super rock-solid technical skills and had a real passion. He really wanted to work on infrastructure at scale. We brought him in and immediately he got working on our metadata caching tier and wrote the first iteration of that which is pretty successful.

Over the summer, Julie Pitt brought in a guy named Chris Brand, who also came from Lawrence Livermore Labs. As I said, it was a goldmine. He, again, had super strong CS skills, was a natural architect. He built this server called the Viewing History Server that took our primitive database-oriented concurrency tracking system for how many people had active streams and he wrote a server for it that was stateful, highly scalable, and worked.

Within the first three months he had put off the first iteration together, put it into production, and it worked flawlessly. It doesn't happen that often. It makes me think how we sometimes award people for heroics when they deploy something into production and it breaks and they save the day. I like people like this, who do it right the first time and, for me, that's an example of somebody who's really thinking about really at the top of their game.

Greg Orzell brought in our first SRE DevOps engineer, a gentleman named Bruce Wong. Bruce [Wong] came from Adobe, was working on productivity tools, and he was a natural recruiter. He would go in and he could get people excited about almost any role and just reel them in. I don't know how he does it still to this day. He was so successful that when Greg [Orzell] moved onto a startup, I promoted him to manage the team and then he came to me and said, "We need to build a device lab. Nobody's doing a centralized support for the device lab. It's hard for engineers to go get everything set up trying to do network testing using Wireshark, getting access to vices, and simulating other network conditions." He pitched for it and we built it. It turned out to be a pretty successful effort.

Finally, in the last year of those four years when I was managing the playback services team, we hired a gentleman named Mohit Vora who was working on our brand new initiative to build the Netflix CDM and he wrote the first version of the control plane. This was essentially the cloud-based service tier that manages all of the caches both in internet exchanges, embedded caches, and ISP networks, etc.

This ultimately laid the foundation for Netflix to become a global company. In 2016, my last year at Netflix, I was no longer managing that team but they went onto build out that foundation to a point where Netflix could launch in pretty much the entire world minus I believe it was China, Iran, and North Korea. That's 136 countries all at once lit up flawless.

Over those four years we accomplished a lot. We went from six high-performing engineers to 40. We were on every device type that matters, thousands of device types. Every digital television out there that had WiFi where we could get a reasonable experience would have Netflix on it. Even put Netflix on a chip that could be embedded in the devices themselves to unique identifiers and provide a lot of the software foundation.

In terms of features and security, the list is so long but examples like the profiles feature, device authentication evolved dramatically over time to be more secure and uniquely identified devices so people couldn't spoof them. There was the encrypted media extensions that Netflix advocated for, created standards around, and had integrated into every browser so that we could do DRM interactions in HTML5. That's a big example of a big push.

Internationalization - we had launched in Canada, Latin America, and Europe. We finally did build that alternate audio and subtitles feature for Canada so we could support French and, of course, that got expanded from there. We built a highly scalable system that could deal with all the permutations and management of regional catalogs, multiple bit rates, a whole variety. Multiple assets, language assets, etc, for every plan.

Then, cloud migration, we did it in record time. It took two years from start to finish for us to do cloud migration. The rest of the company was still catching up after that for a few years, but we were able to do that fairly quickly.

Then, finally, from a technical operations perspective, because we had early focus, because we had people that were so dedicated, we set the standard within Netflix for highly available services, and that's something that we built on and grew over time.

Every one of these people has gone onto do great work in other places, to be high-performance people, to be high-performance leaders building their own high-performance teams over time. Philip [Fisher-Ogden] has remained for all these years at Netflix and now manages both playback and discovery services. Julie [Pitt] left and then came back and is now Director of Engineering for the data science platform. Ranjit [Mavinkurve] went on Workday to become the Director of Engineering there. Greg [Orzell] went through a variety of roles over time and now is at Github where I had hired him and he has stayed on since I left managing, helping them with all kinds of stuff but focusing on Chaos and highly reliable architectures.

The engineers that we brought in who grew over time also continued to grow. Ishaan [Shastri] is finally a manager. He stayed off for all these years and then finally in the last year finally decided to become a manager. Chris Brand is a senior architect at Workday, Ranjit [Mavinkurve] hired him. Bruce [Wong] is now the Director of Engineering at Stitch Fix before being at Twilio. Mohit [Vora] now runs the overall CDN Control Plane team. All of them are virtuosos now, all of them have mastered their craft and are helping to grow and do all the same things that we did during that very important time in 2009 to 2013.

Your Story

I just want to take a couple of minutes and do a little bit of a recap, and then I'm going to give you some homework. I can't make you do it, I'm not going to know if you did it, but I have some things for you to think about to try to really internalize a lot of the things that we've been talking about.

As a reminder, ownership oriented culture seeks to foster a strong sense of ownership in every employee by igniting passion, fostering grace, and demanding fire. We do this by aligning passion with purpose, providing opportunities to autonomously master new skills, and embracing accountability.

When I think about all of you I am not sure who's a manager and who's not. Since we are all individuals and some of us might be leaders and some us might think like leaders even if we're not managers, I'm going to break this into two parts. As an individual, here are questions for you to ask yourself and to think about. What's your origin story and how has that affected your journey? Very important for identifying yourself so you can identify it in other people. What sparks passion for you? Do you wake up on most days excited about what's ahead? If the answer's no, you might have something to think about.

Is your company's mission in alignment with your passion, goals, and values? Does your work provide sufficient transparency for you to feel a sense of ownership and demonstrate your ability to deliver something that matters? Does your work provide opportunities to autonomously master new skills? If the answer is yes or no, but especially if it's no, what can you do to maximize your own passion, grace, and fire? If the opportunities are too limited, think about whether or not you're in the right place or whether or not you will grow in a better place someplace else.

As leaders, we need to then be thinking about others. We need to be thinking about providing these same kinds of opportunities and setting this context for the people on our teams. The real question is, do you have a clearly defined mission and culture so people can opt-in and out with confidence knowing what they're getting themselves into, so they can try to align their passion with what you're doing?

Do you have a clear roadmap and do you know what you have and what you still need looking out? If you don't know what you need right now, do you have a plan to get there? Is your team behind the mission, compatible with your culture, and onboard with your roadmap? If you don't know the answer you should find out because you may not be getting the best work out of the people that you have on your team. Do you practice those things that we talked about around context, not control? Stream transparency, deep delegation, radical honesty, and true accountability.

Is every member of your team a mature fully formed adult? We talked about that earlier. What opportunities for opportunity and mastery can you provide to the team members that align with the work that needs to be done? As a manager, that's probably one of the most critical things that you can do. If your company is not ownership-oriented because the environment matters so much, what can you do to affect that?


See more presentations with transcripts


Recorded at:

Feb 10, 2020