Facilitating the Spread of Knowledge and Innovation in Professional Software Development

Write for InfoQ


Choose your language

InfoQ Homepage Presentations Multiplying Engineering Productivity in Face of Constant Change

Multiplying Engineering Productivity in Face of Constant Change



Shweta Saraf discusses harnessing the collective intelligence of a team to not only multiply productivity, but also cultivate organizational resilience in the face of unceasing changes.


Shweta Saraf is the Director of Platform Networking at Netflix. Prior to Netflix, Shweta was Head of Network Modernization at Equinix. She previously scaled the Edge Infrastructure teams that built the Equinix Metal. She has also held leadership roles at Packet, DigitalOcean and Cisco. Shweta also serves on the board of eBPF Foundation.

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.


Saraf: I've been doing this for 16-plus years, building platforms and products, and been doing it in different environments. I've been working in different environments in the sense of startups, which are 100 or less people, going through IPO, startups 100 or less people, getting acquired by a big company. Then, working at hyperscalers, or cloud providers in terms of navigating tens of thousands of people in a company. My mission in life is really to create the best environment where people can do their best work. That's really what I thrive at. Those are some of the companies I've worked at. Netflix is where I work at right now.

Change Is Constant

Change is constant. Let's just rewind three years back of our lives, what have we faced? Pandemic, war, return to work, macroeconomic conditions, the Great Resignation, and now generational AI, which is everywhere, everyone wants to use it. Changes are hitting us by the wave, in this world. Even as a leader, or an individual contributor, if you try to zoom it down to your daily life, change in our workplaces is very common. It's always happening. Either you have shifting priorities. You end up changing your teams, because people are leaving, or people are laid off. Leaders change, every new leader brings in a new change. Even if things are going smooth, then you might be facing lack of prioritization, lack of clarity, lack of a strategy, or there might be decisions that come in hot, or you might be facing big decisions, like change the language we use, or rearchitect the whole thing. Change is everywhere.


What are we going to talk about? We're going to talk about a bunch of frameworks that I have fine-tuned with my experience coming from these diverse backgrounds. I'm going to tell you more about how I have applied them in real life. Hopefully, that's something you can walk away with which you can apply them in your context. I'm going to build on that a bit.

Measuring Organizational Efficiency

When we talk about change, and what we care about is this whole thing called efficiency or organizational efficiency. No one wants to talk about that, but we all have to as leaders, or fix it. Really, the question that we are asking at the core of it is, how do I get maximum output, the best result I can by investing the least amount of time and resources? The way I decode this equation is really, there's the efficiency part of it, there's the effectiveness part of it. If you have both sides of the coin, that's when you start unlocking organizational excellence, or productivity, as we call it. Let's dive deeper. What do I mean by that? Efficiency is everything that you can measure today: metrics, KPIs, inch metrics, financial metrics. How good are our systems? Can we deliver code to production faster? What does our supply chain look like? Things like that. In my book, accountability is a big one. At Netflix, we are all about freedom and responsibility. It's really interesting how we translate that responsibility to accountability and get stuff done. How do you build this clear accountability with individuals and make sure people are delivering? Communication, if you're not communicating effectively, there could be loss of information, confused people, cascading messages getting lost, all kinds of things. There's also the part where you want to grow your people, grow the talent in your company and get the most out of humans that you have. At the center of all of this is leadership. Does leadership do their part? Are they held accountable? How do they measure up in terms of efficiency? We know all these things. We know how to measure them. Every company has a different way of doing it.

Then, what about this? What about effectiveness? To me, effectiveness is a lens that you can apply to any efficiency metric. Let me tell you how and why. To me it's about, are your teams, is your organization doing the right thing? What is the right thing? It can be subjective. Let's say I have a business, I'm selling cupcakes. If my goal is, I want to sell as many cupcakes as possible, that's fine. That's effective. I can measure it. I can put a metric around it. If I set out a goal, saying, I want to sell the largest amount of cupcakes in the whole state of California, that may not be an effective thing to do if I'm just starting out. I think the measure of effectiveness and thought behind that is very important. I'm going to quote Will Larson, who said this, "It's easy to do the right thing." Think about it, you always end up doing the right thing for your family, no matter how hard it is. It is easy to do the right thing when it's cheap, when you don't have to pay the price for it. It's really hard when it's difficult, and that's when it matters the most. Bringing it back to the example of a workplace. The story goes this way. There was an incident, it caused an outage. This was all before the times of blameless culture and good practices of incident management. The senior leadership asked, who can we hold accountable? Who can we take out based on this? The person who ever did that maybe was a low performer, but it wasn't a thing where doing this action will indicate a message to the org that we are punishing these things versus learning from them. The leader of the org stepped up and said, let me go instead, because that's the right thing to do. I want to create the right culture, and the company culture is not set up to grow people, but punish them for an incident which lasted for 17 hours. That's just an example.

Then, now that we start thinking, ok, this is the measure of efficiency and effectiveness. How do we know if we are succeeding? My teams ask me this all the time, what does good look like? How do we go beyond hitting the numbers, beyond the OKRs? More importantly, what kind of org do we end up becoming, one to two years from now? Even though the company has culture and values, how do behaviors evolve as part of that? I'm going to walk you through some of the tenants that I believe in, in terms of what successful orgs look like. People think success is a measure of progress, it's almost like a linear curve. It's what takes you from point A to point B. In my experience, people don't see this path all the time, but that's how it really looks like. It's all messed up. It's all over the place, but somehow you end up reaching where you reach. Coming back to the tenants, how do I know if the org is succeeding? What do I look at beyond OKRs? The first one for me is, are we rowing in the right direction? A simple test of this is if I'm walking the hallways, or if I grab an engineer on the meet, do they directionally have the same answer to the question of why are we doing something? What are we doing? How are we doing it? Do they have a sense of pride when they say that. If that happens, it means my job as a leader to establishing strategy, clarity, communicating is happening well. It's such a small thing anybody could do and figure out very quickly if your org is indeed rowing in the same direction. The next one is very important to me, because customers are at the heart of what we do. I often see teams, especially platform teams, draw a line around, I built my service. I have my SLA. It's in production, done. No, what's your adoption? Who's using it? Start thinking about, who are you enabling? If you're a platform team, think about your internal customers. If you're a product team, think about your end customers, and get very obsessed with the outcomes. Is the org able to place big bets which are aligned with business outcomes? Knowing that some of these bets will fail, knowing that there is no fear of failure. When you have an environment like that, to me, that's success.

These are the four more things. The next one was high frequency feedback. I don't like annual surveys. Do we have a pulse on our customers? Customers, sometimes in a big company environment are forced to use your products. Not that they love them, they have to. To me the measure is, is my most sophisticated customer, partner trusting our org with solving the most difficult problem they have? An example of this is my team built a service. We moved to 70% adoption at Netflix, for the apps that are on paved road. Yes, it was difficult, but not impossible. Then we hit a point where we started working with our most sophisticated partners, and their use cases were totally different. We didn't anticipate them, there were gaps. Now an approach will be you opt out, you're not part of this paved road. Instead, we said, no, let's take a pause. Let's collaborate. Let's fix those gaps, get to that 100%. In hindsight, we should have frontloaded these problems, we should have eaten the frog. That's the approach, versus people just are forced to use the stuff that you're building. Execution excellence, say what you're going to do, and then do it. That builds credibility versus anything else in my book. Then, this is an interesting one, especially for platform teams, we talk about leverage a lot. The dilemma I see they fall into is, are your platform teams doing the work that other teams were supposed to do, just shifting the responsibilities, or are they doing it in a way where you create leverage, where you multiply? To me, that's where it's like, are we selling our ideas? Are we trying to get the buy-in, or are we creating value where people want to adopt it? Tied to the previous thing. The last thing is something I try to do as a leader, I want to work myself out of the job that I have, in the sense that, am I investing in growth? Am I creating the leaders of the future, so that they can run this team or organization better than I can so I can go do something bigger or different? These are the things that I look for to say, are we successful?

The Multiplier Framework

Now we're going to talk about some of the frameworks. The first framework I'm going to talk about is multiplier framework. I didn't invent this. Liz Wiseman wrote a book. I highly recommend you guys read it. The reason I'm going to talk about this is we established how you measure org efficiency. We established what success looks like. Now I'm going to tell you, how do I bring in orgs to a level where they are successful. One of the aha moments was, many years back even before I knew this was a framework or this is what it was called. It's not about hiring highly talented engineers. You need that high talent density. It's not about infrastructure and tools. You need that, yes. What happens? That's not enough. You need to figure out how do you get collective intelligence harnessed from your teams. Just having smart people on your team is not enough, you need to start multiplying. How many of you have worked with horrible bosses, or extreme diminishers, any of these personas? This was the aha moment, actually. People normally grow into this thinking, "I as a leader need to be in the middle of this decision. People can't figure it out without me. I'm playing an important role." The light bulb is really, no, people are smart. You hired them because they're smart. Get out of the way, empower them. People say empower them, but what really that means. They will figure it out without you. You should be ok with that. The survey shows that people who have diminishing tendencies, leaders who have diminishing tendencies, can only harness less than 50% of total potential impact on their orgs. People who have multiplying tendencies can do double of that. Let's dive into it.

What does a multiplier look like? It's not like they have all of these, they may have one or more of these. This is what multiplying looks like. People want to work for you, you're a talent magnet. You create a liberating environment. You create debate, which means you debate and then decide, not decide and then debate. Then, you instill a sense of ownership, accountability. On the other hand, this is something you all might resonate with, going back to horrible bosses, they might have one or more of these tendencies, or all of them if they are really horrible. Basically, empire building. Know-it-all, people can't figure it out without me. The worst is micromanager, trying to breath over people's neck telling them what to do. Where do we land with this? Our goal is to amplify multiplier qualities, remove the diminishing qualities, and what we get is the collective intelligence on our teams. That was really the aha moment.

The other aha moment is nobody wakes up and says, I'm going to go diminish someone. Most of the leaders identify themselves as accidental diminishers. The tool that you need in your toolkit is identifying what are your accidental diminisher tendencies, and then figuring out how to deal with them. I'm going to talk about two examples. Idea fountain, I was guilty of that. I have a lot of ideas. I come from a place of enthusiasm and passion. That ends up overwhelming the team and maybe not having space. I had that aha moment many years back. Since then, I've worked on it. The other one I see is rapid responder. People who respond to Slack, email, or whatever, any single minute that you reach out to them. Setting an expectation that it's helping the org move faster, but they are really reacting to a situation, they're not responding. They are taking on a lot of parallel threads, which actually creates a traffic jam for the team. Bottom line is, even if you have best intentions, they may not lead to the best outcomes. That's because people might be accidentally diminishing.

What I did with my team was, and I've done this with multiple teams, and it has worked every single time, we sat down and we talked about, like describe this framework, talk about what are our accidental diminishing qualities. I started first. When I said idea fountain is one of my primary one, it created a safe space for my team to then open up. People were like, "Yes, I think I'm always on, or I'm a rapid responder." It created a common framework where people got vulnerable, people started accepting and being ok with this kind of language. Then we picked one tendency, we ran an experiment for 30 days. It became ok for the team to acknowledge where they need help. Whenever such things happen, use this framework to redirect versus assume that they have bad intent. We used to start off some of our team meetings by sharing like, this week, I had a multiplying moment by doing x. This week, I had a diminishing moment by doing y. The results are, once you start multiplying, you start seeing the empowerment in action.

Adapting to Constant Change

Now I'm going to talk about change, because that's a big part of this topic. How do you adapt to change? The second framework I want to give you is the change mindset model. How do you go through change? I'm going to give you an example here as well. At one of the companies, the startup I worked for where I led engineering, got acquired. A simple thing like the startup used Slack, the company used Teams, big debate. No one wants to agree which way to go. Slack, obviously, but still? I'm purposely taking a simple light example. There were many other such things. The first bucket or the mindset is that of a resistor. You have met these people, any change you want to do or a change is happening, first reaction is no. It comes from a place of, they want to gain authority, they want to push back, they want to create a bubble or protect against chaos. The important thing is not to just brand this as a bad thing, but to understand this mindset as a leader, because you can learn from, what are the things they are foreseeing which is making them resist the change? How can you lean in as a leader to push through that? It reminds me of Yennefer from Witcher. This is where it's another common phenomenon. Many people fall into this bucket, change receiver. Change is happening to me. I have no influence on this. Just tell me what to do and I'll do it. I don't want to have conflict. I don't want to get in the way. Nothing bad with it.

Change controller, these are the people who plan. They plan. They want to drive change. They want to control the change. They want to control the narrative. Reminds me of Davey who plans but things don't always go through as planned. This is another mindset. Last one is change ready. Many leaders operate in this mindset where they focus on aligning people on the why. They know change is not a one-step journey, especially if the change is multi-quarter, or multi-year change, multi-dimensional change. They co-create the hub. This is how I think about it. In that example, the justification was all our customers are using Slack, we have shared Slack channels, 10,000 people in the company are using Teams, we don't want to move away from Teams. Initially, it became like, one versus the other. As we did analysis, as we understood feedback from these different mindsets, we came to a conclusion that engineering will primarily start using Slack. We'll still keep everything else for everybody else, finance and so on. Two things will coexist, versus killing one versus the other. Maybe a cop-out, but worked out.

Again, what I want to leave you with, this is Beckhard-Harris model, and not too much aligned on the math or equation, but I think this works. Every time I walk into a new team, or in a new scenario, people are not satisfied with tooling, with process, with migrations. If there's dissatisfaction, and then you take the first step as a leader to create a compelling shared vision, that's where you have to build consensus by understanding those different mindsets and overcoming them. The last ingredient is very important, like have the first step, which is a creator, which is action oriented, then you can overcome the resistance to change. That's the key for driving long-term change or even receiving the change, that kind of mindset.

We spoke about change. One last thing I want to talk about before I dive into examples, is something I talked with another fellow leader of mine sometime, is wartime versus peacetime. Andreessen Horowitz had written a beautiful article a decade ago around wartime CEOs and peacetime CEOs. To me, it's more of the mindset. It's not like, yes, there's a part of nature. Naturally, you might incline towards one versus the other. Oftentimes, you're put into a situation where it feels like you're in wartime, so can you adapt? As a leader, can you change? Just to narrow it down, I feel when you're in wartime, you do have to think about short-term iterative wins versus trying to build a strategy for long-term investment. You may not have the luxury of doing that. In wartime, you may have to do all hands on deck. In peacetime, you may have teams with separate lanes where they are highly aligned, but loosely coupled. In wartime, you have to rely a lot on action and instinct. In peacetime, you can build consensus.

Example 1: DigitalOcean

Let's get into the examples. Now that we talked about efficiency, effectiveness, we talked about multiplier effect, we talked about change mindset, and wartime, peacetime. Those are the three go-to frameworks of mine. I'm going to show you how I applied them in real life through different scenarios. First one, just for context, DigitalOcean, a cloud startup, when I joined, less than 100 people. We were growth hacking after series C. We were going up against AWS, GCP. Don't look at the stock price now, but at that point, we were very successful in doing so. The reason was we focused on simplicity and developers. We were able to zero down on few things, do them well, versus go behind the full feature set that clouds provided. Specifically, what was the impact? The impact was bootstrapping an entire networking portfolio that I led, starting from zero to one, and then scaling it up. Delivering a strategy and roadmap which was actually one of the key ingredients for us to go to IPO, because we didn't have a strong network story when I joined. We had a compute story and a storage story, but network was essential for us to go IPO. That was the scenario where we were in 2016. The approach I took, I had to take wartime leadership approach, because we were not number one in the market. We didn't have thousands of engineers. We were very nimble. We had to make it work. We had funding. We had customers, so we had to really survive. We had to fight it out. I also took the approach of change controller mindset, because this was a situation where there was no status quo. People didn't know what good looks like. People were not opinionated, so they wanted somebody to come in and say, "Let's do it this way. This works. Let's go with this approach."

Then specifically, some of the things I had to try was actually build versus buy strategy, establish that. Because we want to create differentiated value for our core offerings versus rely on buy, because as a startup, it's easy to buy, but we had to make hard decisions to go build. We had to hire a dream team. We had to multiply. We had to do all this in parallel. That approach worked out really well because when we did it for about two-and-a-half years, we were able to deliver the entire essential cloud portfolio, which was VPC, load balancers, and firewalls. We ended up going IPO. The challenges we did face were, we had to keep moving fast, we didn't have time to think long term so much. We had IPO timelines, so there was a definite time-driven goal. Then convincing the CEO. If you're a networking person, you either have a CEO who gets it or who doesn't. If they don't, then you have to just keep convincing them about what the right decision to do is. They don't know that. This is just an example of how those strategies or frameworks came into play.

Example 2: Packet

The next one I want to talk about is another startup I did after DigitalOcean, which was Packet. Again, joined when the team was less than 100 people, pandemic hit. Right in the middle of the pandemic, we were getting acquired by a giant called Equinix. It was tough because you had to lead through unknowns. You had to lead a team through acquisition. You had to do rebranding of all your product lines. Then you have to go scale it, like literally scale it from 8 regions to 35 regions overnight, almost. Then you had to hire a global team and keep the morale high in spite of everything. The other challenge was, we really operated a whole stack so the diversity of skill set was real. We had to hire people who were good at hardware systems, network, product UI. This was a phase of high growth and high learning, high toil as well. Impact wise, as I said, some of those numbers, the team 5x'd. The footprint grew really huge, and revenue grew really huge. We went from delivering one product, one platform, to multi products, multi-platform, which was a shift in how we think and how we develop.

What mindset did I have to adopt here? I had to adopt a change-ready mindset. Again, wartime leadership. It really felt like that little tugboat trying to move that whole giant thing. It was not easy. The whole motto was, we want to deliver hardware at software speed. You can only do that by bringing in agility and speed of delivery. To make a giant company who is not used to operating that way, it's a huge change in the mindset. It was not only, hit your numbers, deliver on your OKRs, get the startup acquisition successful, but also attempt to move this huge ship. It required a lot of leading by example. It required a lot of challenging the mindset, but also willing to do the work to bring people in. This is context, this is not control, you have to bring people in. Then really, not give up, in spite of all these hurdles. Challenges in this environment was, like I said, it felt like a constant uphill battle. There were days where we felt this is pointless, but we kept going. Then, going fast was not the optimizer, going together was the optimizer. The work to bring people in was a heavy tax versus just hit your goals. As a leader, these kinds of frameworks helped me navigate those situations. I had to create new markets for the company where the company was not comfortable operating. By doing the work, leading by example, we were able to grab market, which was net-new to the company in terms of bare metal cloud. That is the second example.

Example 3: Netflix

Now, third example. I've been here only for five months at Netflix. My team is responsible for all the traffic that flows through our networks when you try to go to the app and select which title you want to play. A lot of tier zero services, a lot of critical responsibilities. It's still work in progress. I don't want to say the work is done. Some of the things I've been able to do since I joined is establish clear accountability. We talked about that in the efficiency. No blurry lines. People were not able to move forward because the decisions were not made, or people were waiting on each other. Strategy for next 18 to 24 months. We are huge on strategy at Netflix. As a leader, it's your responsibility to really understand partners, customers, and identify what are the problems. Then go make bets, because these are hypotheses. This is how you might be able to bet, and with data, fulfill them, execute on them in next 12 to 18 months. Then deliver the impact. Then, continue to hire the dream team. This is something we have done. My approach here was, again, change ready and multiplier, because change is constant and Netflix goes through constant change for those of you who are in the company. My approach was to establish trust, and establish accountability charts.

I had to partner with PM and unify the strategy in the org. I had to do some tough decisions in terms of changing the org. Challenges were really around, we operate at huge scale, so migrating anything as a platform team is a huge challenge. While we are good at hiring, onboarding people is a huge challenge. Prioritizing multiple work streams, with multiple teams we support is a huge challenge. Having the change-ready mindset and multiplier approach aligns with our values, where we are all about context over control. We want to empower our teams because we have some extremely smart engineers here working at Netflix. My role has been really to strengthen the org in the sense of what it needs and remove what it doesn't need. Hopefully, you got a flavor of how you can apply these frameworks, in your context, in your problems, in your organization.

Expanding Roles

The last thing I want to talk about is, great, you mastered all these frameworks. You're doing great at leadership. Your org is thriving. You're successful. Awesome. We change the rules of the game now, now what? When does that happen? It can happen. When do you really get into expanding roles or changing roles? I want to talk a little bit about, you're applying all these tools, and you're in an environment and in a known quantity, but what happens when the rules of the game change? Which is like, now I'm doing a new role, I got promoted. Or I'm taking a new role up, or I'm joining a net-new company, how do I navigate it? What do I do? I just want to give you a little bit of how do I approach it, and how do I apply all those mechanisms, but also create a plan in terms of what really matters. I think the biggest thing is, if you're joining a new company, you need to grok the context, the culture of the company, very quickly. That will determine your success, because you may think you know, you may think this works, but as the rules of the game change, you will have to adapt and change as a leader. The most important bit is, don't wait. Hit the ground running. Hit some quick wins. Create value for the company. At the same time, don't do it from a place where you know it all, or this worked for you in company x, so it's going to work here. Really spend time to understand the context. Really spend time to earn the trust. Those are going to go a long way. At the same time, once you figure that out, do not wait and see. Say what you're going to do, and then do it. That's the mechanism of creating credibility.

Becoming a Multiplier

I just want to leave you guys with, you can start multiplying today. We talked about some techniques, which are easy to apply. You can do it. Who's ready to multiply with me, and this hero of Stranger Things?


We talked about my understanding and framework of how to measure efficiency, effectiveness. How can that unlock productivity? What is multiplier framework? How to identify accidental diminishers in your context, in your team, even starting with you yourself. How to understand and apply a change mindset model and use it to your advantage, versus get frustrated by people who are not ready to change or are too quick to change. Iterate as your role expands and changes, and establish the new rules of the game very quickly if you're in a new org. Then, most importantly, you can start multiplying today. You don't need anything. You can start doing that today.

Questions and Answers

Participant 1: Any tips on working with leaders who don't share the same value set as you do, because you have to work with leaders that are at the same level? Also, sometimes how do you work with somebody whose values are completely different from you as a leader?

Saraf: I think the right answer is if there's a value misalignment, you really need to ask yourself, why are you here? If it is more of a behavioral or priority misalignment, the practical answer is, figure out what they really care about, and try to see how you can meet them in the middle. Because you could be very much idealistic, but that doesn't work in a regular environment. If there's a core value mismatch, my approach would be, I would question, why am I here?

Participant 2: Can you share a little bit more about that accidental diminisher. There was a slide with a lot of different qualities. Some of them I was really curious about. Like, optimist, protector, strategist, perfectionist. Just a little bit more detail on those qualities.

Saraf: Protector, for example. There are leaders who do not share context with their teams. They think they are the poop umbrella. They need to protect them from office politics or things like that, and there's a balance. Then, if you do too much of that, you don't let your people fend for yourself, and you end up accidentally diminishing. If you're an optimist, you might come across as, yes, we can do it. We can do this. We can do that. Then people might end up feeling you don't grok the nuance, you don't understand the difficulty in this. Even though you're coming from a good place, your engineers or your team might end up feeling, does this person really understand how tough it is? Do they have an ear to listen to you? Pacesetter, rapid responder is kind of similar. In a way pacesetter is, you always go-go-go. You're trying to get your team moving, but it creates an unhealthy expectation that I have to answer Slack, or I have to keep going all the time, even if that's not what you think about. You care about the outcomes. If you show up in a way where you're setting the pace, your team is going to start following suit. Always on is another one we talk about a lot. There are people in meetings who take up all the space, and they're doing it because they have passion, or they have things to contribute, but they're not mindful that this is taking away time from someone who's shy or who also has ideas but will not speak unless prompted. There are experiments that you can do to say you have only two coins, which means that you have to wait until three people have spoken to use your coin or something like that. That's how it goes. You can look it up in the book and there's a whole theory on that.


See more presentations with transcripts


Recorded at:

Jun 05, 2024