Season 8 - Episode 8
Building a Product Culture with Twitter VP of Engineering
As VP of Engineering for Twitter, Nick Caldwell has experience in general management, engineering, and product. In this episode, he dives into the intersection of these roles, how to build a strong product culture, and dealing with the unique challenges in social media spaces.
“How did you break into product? ”
I’ll go way back in the day. I’ve always been passionate about technology in general. I grew up in PG County, Maryland, which is a 90+ percent black suburb to the Southeast of DC. My dad was a public defender and he used to do all of his work on a typewriter. One day he brought home a Tandy 1000, one of the first personal computers. So I used to sit on his lap and pretend-type. Fortunately, I didn’t break anything. That led to an early fascination with computers, starting with gaming like playing Wolf 3D, and Doom. I got involved in the early internet scene and bulletin board systems.
Those gave me access to a wide variety of people and opportunities that I wouldn’t have otherwise been exposed to. It went from there. I got into ML, AI and taught myself C++ and so on and so forth. It took me far away from PG County, Maryland to school at MIT and then to a job on the other coast at Microsoft in Seattle. I waited way too long in that role and eventually made my way to the bay area startup scene, which has been absolutely fascinating in the past six years that my wife and I have moved down here.
“Being a general manager at Microsoft and then taking roles in both product and engineering, what is the intersection between them? ”
I was very fortunate in my career at Microsoft. For those of all who aren’t familiar with Microsoft, it is very heavy on the general management role as you climb up the ranks. The higher you get at that company, the more exposed you are to different aspects of the business to the point where you are asked to inherit and run teams, and eventually, own entire business lines. While I was at Microsoft and in other roles, I’ve run product engineering, design, product marketing, primarily focused on the R&D side of the house and less on the go-to-market side of the house.
Both roles, product and engineering, care very much about the product and what’s being built. Both help to have an understanding of technology and what’s possible. Both care about people as well. I think the differences are important. It comes down to what sort of people each of these groups cares about and why it matters to the success of the overall team. So, both of these groups are very important in different constituents.
For product folks, their constituents are customers, all the cross-functional teams that it takes to go into product creation, the sales team, the interface with the go-to-market. You’re dealing with a very wide variety of functions, customers, etc, and trying to synthesize that into product strategy, product roadmap, and so on and so forth. Engineering leaders also have a large number of constituents, but they’re primarily engineers. Being able to cat herd large numbers of engineers is a skill in and of itself.
There’s an intersection there, engineers, like product people, need vision and motivation. They need to understand the mission and the why of what they’re doing. But like any other function, engineering also has its own way of working. When you’re in engineering, you’re not working in a factory line. It’s not like the job is to sit there and put beans in a can. It isn’t a very straightforward robotic exercise. It’s much more like craftsmanship, but it has to be done at scale with predictability and quality. As an engineering manager, you’re trying to coerce all of these teams of craftspeople into building on a schedule to whatever the broader vision is. So managing that constituency becomes your primary job.
“It’s refreshing to see that you were able to understand both sides of the equation and jump in between them. I’ve seen situations where product tends to be more business driven and assumes that engineering is just executing what product says or engineering thinks that they’re receiving orders from people who actually don’t know how to build stuff. So getting both perspectives is valuable for the entire team. ”
That’s the worst. When the disciplines aren’t talking to each other, you don’t build the best product, unless you pull in all the different opinions and insights to figure out a way to best synthesize them. I would say, on the whole, a product person who doesn’t know how to do that synthesis is not going to be very effective. You have to learn how to talk to each of the stakeholders in a way that aligns with not just how they like to be spoken to, but that aligns with their goals and their outcomes in OKRs. Marketing people, engineers, salespeople are literally gold on different things. So as the product person sitting between all of these folks, it really helps and is important to understand what their motivations are before you go talk to them and try and get them bought into whatever strategy or mission is.
“Could you tell us more about building a product culture? ”
Any company you drop into is going to have culture and leadership values that are indirectly or influenced by their customer base and whatever their product is and their history. That is what culture is. Culture is essentially the accumulation over time of what behaviors you reward and punish. I guess that’s the textbook definition of it. So Microsoft has a long history of being very competitive. I don’t even know how to describe Twitter’s culture. For any company that works in social media, social media companies tend to be about authenticity. So I worked at Reddit and Twitter as well. Internally, we highly value and reward people who are authentic and bring their true selves into the workplace.
I also spent some time at Google. When you join, they give you a propeller head hat. It’s much more of a technocratic culture over there. The main thing though is that any of these things can work. It depends on what your product environment is and being open to iterating your culture to match whatever your present-day challenges might happen to be. That leaves a lot of room for interpretation. If I had to give you some guidance on how to build a product culture that works well, if you need a starting place, it’s encouraging people to be fearless. So not worrying about taking risks, being okay going into new areas and seeing what happens, and having accommodation for trying new things.
I don’t want to say failure. I think I actually disagree with just being okay with failure. You should have a success plan in mind and if it fails, you iterate. But don’t aim for failure, aim for success. Allowing people not to have artificial boundaries. In my mind, there are a lot of artificial boundaries when it comes to building products or the way that we work. A lot of these boundaries are put up to give us structure and predictability and familiarity, but they can also be limiting factors.
We talked earlier about switching between product and engineering. Imagine if I had put a boundary on myself like since I’m an engineer, I’m never going to care about product. My career would have been nowhere near as successful as it has been. So, now apply that to entire organizations and say “limit what you can think about” or “limit how much impact you can have on a particular business line that might already exist”, you’re just narrowing the scope of what’s possible. So you should not place too many artificial boundaries on your team.
The other thing is that having a genuine appreciation for teamwork makes a lot of things easier. There are cultures that you can work in where you’re more motivated toward individual performance, or maybe even more rewarded for individual performance. Thinking about systems and putting reward mechanics in place that elevate teams and allow people to succeed together can have some really outsized outcomes in a lot of different dimensions, including better team cohesion, more creativity, and people filling gaps in terms of supporting each other.
All these boundaries we put up are usually artificial. They may be there for a reason like to correct some mistake in the past, but you should always question why the boundary is there. Is it something that I can overcome? Oftentimes the answer is yes. Just remember not to put artificial limits on yourself, it causes lots of downstream problems.
“ What are some of the big problems that you are passionate about solving? ”
I think career-wise, there are two things I found most interesting. I’ve spent most of my career working in data and data engineering and BI space. My big thing at Microsoft was this product called power BI. Then fast forward several years later, I was the chief product officer at a company called Looker, which was acquired by Google.
In both of those cases, the thing that made us successful was that we were trying to make data available in an operational way. That is to say that even though data can be challenging and you can do all this crazy cool forecasting and analysis and machine learning, ultimately none of this stuff matters unless a normal person can get something done with it.
Normal doesn’t mean having to go learn SQL. Normal people are like school teachers or plumbers. Both in power BI and Looker, we strove to build a data platform that would collect all the data, make it simple, but ultimately allow you to build applications and services for everyday people. And that is what allowed us to be successful. So that’s one thing. I think that’s still a very important trend in the market right now.
The other side of my career has been in social media like Reddit and Twitter. There are so many product challenges and societal challenges in social media. A pithy way to say it is that large numbers of people kind of suck and trying to figure out how to build products that keep large numbers of people interacting together in healthy, safe, and positive ways that are good for the world is a fundamentally hard problem that all social media companies struggle with in some way shape or form. So, that’s something that’s been pretty exciting for me both, at Reddit and Twitter.
“In a way, both things that you mentioned connect because the bigger those communities become on social media, the harder it is to say something that is going to be liked by everyone. So the relevance of data to understand how to target specific groups to make sure that they feel engaged and not to overwhelm them with other content out there must be incredible. ”
There’s a lot of work that goes into relevance and personalization, but that can also have some counter effects where you can create echo chambers and people start amplifying ideas that aren’t good. So it’s a very challenging problem. If you think about every couple of generations, we get a transformative change in how we communicate with one another, like printing, press, radio, television, etc all the way up to social media. And every one of these advancements has allowed individuals to reach a larger number of people more quickly. Each one has been sort of followed by turmoil and upheaval. Giving people access to these technologies that allow them to extend their reach often immediately has consequences that need to be sorted through.
“As a leader, you have to think about allocating resources to nurturing your core business while at the same time innovating in other areas like the social audio space. How are you able to really react, plan ahead and really spin out something that fast while maintaining the core? ”
As a product Visionaire, I predicted the social audio space years ago. This is one of the things I’m really proud of about Twitter. That is to say, in the back half of last year, about a year ago, we saw growth in this space, obviously with Clubhouse. There were lots of different apps, as you said, that started to get in the social audio space. If we were at a bigger, slower-moving company, we might have waited to just see how the whole space shook out.
By the way, I’m not the product guy. I need to give credit to Cavan who’s our actual product guy. So I’m going to be speaking somewhat on his behalf. He saw this trend emerging and was like “look, we have some of the technical pieces around and some of the expertise on the product side lying around from a previous acquisition with Periscope, maybe if we rejigger our organization, we jump into this and make a really big impact quickly.”
He was right. The spaces team at Twitter did not effectively exist in January of this year. We were able to get a core group of people. We dissected the necessary skill sets and folks we would need both on the engineering and go to the market side and we threw it together really quickly., More experienced folks like myself who have done other rapid org designs worked with that team to make sure that they had a strong growth plan.
If you ever worked at a sizable company, you try to bring in third-party technologies like procurement and budgeting, and we cut all that red tape. Then within about two to three months, they had put together an extremely competitive starting point platform that they’ve just been iterating on ever since. With that initial success comes the go-to-market functions and the content partnerships and so on and so forth.
I’m proud of the team. In my career, I’ve shipped a lot of stuff pretty fast. Power BI is now a multi-billion dollar business. That took us about a year and a half to go from standing start to launch. With Spaces, it was fewer people, fewer resources, and going really fast. So I’m optimistic about that. It also says a lot about the company. We’re willing to take these very big, high beta bets while simultaneously growing the core business.
“I spent most of my career on the startup side and two to three months is actually a very short period of time to ship something. So I can’t imagine what it’s like at a much larger organization if you have to add other layers of complexity and compliance. ”
I have to give credit to the team. All I can do is cut red tape and make it easy, but it takes passionate people who really love the idea. I wish I could have invited some of these folks, but if you talk to them, you’ll understand this isn’t like a catch-up with Clubhouse. They really believe in their bones that in the future of social audio, they want to be market leaders. The fact that the overall market tailwinds put some wind in their sails, really only amplified the passion that preexisted. You can’t get any better starting point than that. Smart people, well resourced and market tailwind, I’ll bet on that every day of the week
“How can you maintain a roadmap and keep a balance between allocating the right engineering resources so the initiatives have the right support while also ensuring your core is stable and thriving? ”
This is a challenge for any company because what you’re really talking about is how do we make sure we don’t fall into the innovator’s dilemma trap? Twitter obviously has a really well-functioning core business around tweeting stuff and how much of our investments should be put into these nonlinear bets. There isn’t a one size fit all answer for that. It’s basically a process that occurs in planning. You have to understand how much upside there is in your core business and you also have to be able to weigh any of these high beta potential opportunities that we have. Twitter does this process once a quarter and looks at where our people and resources are allocated across the portfolio of investments that we’re making.
Then we do a best-effort guess on the ROI of any of the investments. When I say ROI, I mean tied to business outcomes, like monetization or DAU. We try to tie it to some overarching business metric. Things like spaces where there isn’t necessarily a direct line towards existing business metrics. we have to play it a little bit fuzzier and that’s okay because that really is a new category. Being willing to operate in that new category with uncertainty is sort of an advantage in and of itself that other companies might not take.
Once a quarter look at your portfolio, is it balanced the right way? People often ask ‘What is the right portfolio balance?’. I think, on the whole, you should obviously be doubling down on things at work. So, as fun as spaces and some of these new investments are, the majority of our investment actually goes toward Twitter as it is. Then we carve out some percentage to tackle these new areas and hopefully, those bets pay off and will double down and put more resourcing into them.
“As you think about your own personal roadmap, what are you curious about learning these days? ”
I guess there are two things I implied. I started my career in machine learning and I started too early. It feels like machine learning now has really taken off with deep learning models. . When I started, you were building neural networks from the ground up, like hand code. Now, there’s ML as a service. I could get off this call and go and get a farm of a hundred tensor machines and start training all sorts of wild things. I would never have imagined this even five years ago.
So I think ML is beyond an inflection point. It’s really becoming amazing. Then the other one we just talked about, something is happening with blockchain and Web 3.0 that I wish I understood more and. I think five years from now, it will be clear what this all means, but there’s something genuinely happening here. It’s easy to be cynical about it because when you go and look at these things, it’s people buying and selling pictures of monkeys for like $300,000 like that. That’s probably not where this market’s gonna end up. I had a chat with Dharmesh from HubSpot and he gave a presentation about his view on web 3.0.
As an example use case. He talked about how decentralized profiles could be used to replace many of the problems that exist with social media companies like LinkedIn, or Twitter owning your profile information. He explained it in a way where I finally understood how I would actually get value if such a service existed. There are other opportunities. I think web 3.0 Is going to potentially cause us to take more dependencies on open protocols for things like chat and direct messaging, which could have broader implications on the market. So there’s something here that is beyond selling pictures of gorillas. I don’t know exactly where it’s going to end up, but it’s worth looking into. Fortunately, I can do my day job where we actually invest in some of these things.
Then I’m on board and advisory roles, which allowed me to participate in some of the discussions as well without having as much skin in the game. I still get to learn about it. I guess that’s a bit of advice I would give to people listening, Don’t try and get too trapped in your local company ecosystem. You should always have some feelers or network into the outside world and learn about other companies or other opportunities or other market trends. In general, the more you can learn the better. Your network, in the long run, is your net worth. Having more access to these opportunities makes you smarter and it could bring more opportunities your way.
“I think that adopting a student mindset is the approach that I personally take and encourage others to do. This is a journey and there’s so many new things happening that it’s easy to get stuck. ”
When you’re in management long enough, you learn that you have to depend on people. Being open to change is a huge thing. Right now, web tech seems to have a new framework every six months. The beauty of this industry is how quickly it changes and you have to have some strategy for how you’re gonna accommodate that.
“What are some of the things you do that don’t scale but are important for you to continue investing time into them? ”
I’ve always been very passionate about management and people development and organizational design. So even to this, my org now is like 750 something people. A bunch of people works for me, but the main thing is I still like to do one-on-one. When people are willing to talk to me, I learn so much from just talking to software engineers. Those are some of my favorite conversations.
Then team meetings as well. I’ll go pretty deep into the org and just have pretty open conversations with people on the team. This doesn’t scale at all. It can be pretty intimidating if it’s not set up in the right way, because I guess if a lot of people work for you, you no longer are treated like a regular person, but I am a regular person. I like meeting people. So, it doesn’t scale, but I learn so much in those conversations. I always tell people in these meetings that all the knowledgeable smart people who know what is actually happening in the org are at the bottom of the org chart. So if you don’t have mechanisms to funnel, funnel that up to leadership to make better decisions and something’s broken. I use these one-on-one and team syncs not just to meet people, but to genuinely change how I think about operating the business.
“If you could give advice to your younger self, who’s thinking about breaking into product, what would you say? ”
There are a couple of things I would say starting with learning to get out of your comfort zone. I spent 15 years at Microsoft. I think I mentioned this earlier, I learned a lot there, but I learned an enormous amount when I finally decided to break out of that ecosystem and I moved to the bay area. I was at Microsoft because I was very comfortable and good at that job. I told you all I grew up in PG County, Maryland.
My parents were lower-middle class. My parents were both public servants. So we had relatively little income. When I got my first job at Microsoft, I was making more than both of them combined. So there was sort of an economic motivation plus my job was good. I stayed in that place for longer than I should have because I had gotten comfortable. I was also a little bit afraid of trying something new. When I worked up the courage through all sorts of mechanisms and was able to jump into something new, I really started to learn a lot. I really started to get access to some great opportunities. Just try and get comfortable being uncomfortable.
The second one is that the job ladder only serves you up to a point. Similar to how I just talked about being locked into one company’s ecosystem, don’t lock your mindset into a job ladder. A job ladder is a one company artifact that helps frame career progression for that company and its culture. If you take two different companies and stack their job ladders up next to each other, there’ll be different things on it. So you need to learn that these are constructs created to help a particular company. You should not wedge yourself in your mentality to that.
A better way to think about navigating your job ladder is to think about all the skills, people, and all of the experiences that you want to have in your career, and then try and navigate to those. So as far as the job ladder lines up with that and helps you, that’s good. But don’t use it as a limitation on things you want to do and opportunities you might pursue. Don’t even use it as a way to value your self-worth. That’s the worst. If you use some company’s job ladder construct to measure your own self-worth, that’s not the way to think about your career advancement.
I guess my final advice is that managers don’t manage your career. You have to remember that managers are managing your workload. It’s great when they can break out of that mold and think about helping you with career opportunities. But, you should spend your time and energy also thinking about how to look for opportunities, how to network, etc. Don’t put so much of that expectation on your manager. You have to manage your own career. So that’s what I would tell young Nick.