E2M with Twitter VP of Engineering Nick Caldwell
“Management didn’t mean sitting around and being told what to do. It meant a better understanding for everyone about the impact of what their day to day work would do. And predictability that would allow our product team and our other go-to-market functions to trust us more.”
Nick Caldwell recently joined Twitter as the VP of engineering where he leads a team of 700+ technical folks who are building Twitter’s consumer-facing products. He started his career at Microsoft where he led Power BI. From there, he jumped to Reddit and he shares a bit about the culture shock experienced going from an organization that’s somewhat resistant to fast, nimble moves to one that needs to be convinced that management isn’t a four letter word. He also dives into the alchemy behind getting consumer product right. In his experience, you can have mountains of data and conduct all the user interviews you’d like, but still need to internalize as an engineer or a product person that a lot of the things you try just aren’t going to work.
Listen now to the full interview:
More of a reader than a listener? Below is an excerpt from Nick’s conversation with Steve Herrod and Quentin Clark for the podcast.
Nick Caldwell | Equivalent to Magic
Steve Herrod: In this episode, we have Nick Caldwell, the VP of Engineering at Twitter. Nick joined Twitter over the summer, after a long career building enterprise products. He now leads a team of 700 engineers on the consumer side – presenting him with a very different set of daily challenges.
Quentin Clark: Nick was hired at Twitter because he’s really good at managing teams and setting product roadmaps. He got his start at Microsoft, where he worked for more than 15 years. I worked with him there when I was leading the data platform organization – SQL server, BI, and such, and he was leading big parts of engineering for some of those products.
This was the 2011 to 2014 era; we had a few very nascent products that we called Azure but really had not committed to transitioning our whole enterprise portfolio over to cloud services. We made that shift over this timeframe – and it was a huge effort. Everything known today as Azure was created in that time. Everything changed – architecture, the business model, culture, how we engaged customers, how we ran our teams. It was messy.
Nick Caldwell: And I just remember it being not just messy, but confusing to try to do all that at the same time because are you prioritizing the end-user product, the infrastructure. Initially, there was a lot of lack of focus.
Steve Herrod: But simultaneously, you were going through a big ramp and the number of people you were managing, I believe. But what was it like kind of as a leader who started with a couple of dozen and then kind of got up over 100 people? What were some of the lessons learned there?
Nick Caldwell: I think we Power BI by the time we’re done was like, 300 people. So it was incredibly fast-growing. And we were growing those teams in Microsoft. The way you grow a team in Microsoft is it’s not like I went external and hired 300 people off the street. At Microsoft, you’re glomming together different teams. So it had this sort of real challenge of how to pull people together into one cohesive culture and rally them behind a single mission. And honestly, we grew so fast…
“I definitely won’t claim that I was able to make this completely cohesive. It was like managing four or five different tribes, each with a slightly different way of working. And what I came away with was a real appreciation for being able to clearly talk about vision and roadmaps and execution at scale.”
That is, with hundreds of people, you have to be able to crystallize what’s important down to like marketing slogans. And get your leadership, your next level, down to repeat those as well and add the right sort of inflection on them, so that’s relevant to their specific team’s interest. And that sort of recurses all the way down the line.
Quentin Clark: So how did you think about which things to really dive into and where to spend your time on the deep technical issues?
Nick Caldwell: I think for me, the high order bit was speed when we were time to market. And there’s a broader context here, which is if we were not successful in figuring out a BI product, and BI strategy that worked, there were other acquisition targets that Microsoft could have gone after. So for us, with that backdrop in mind, we’re like, “Hey, we need to figure out something quick.” And that often meant going outside of the core Microsoft tech stack and leveraging external tools.
I think D3 and Modern Web were probably the biggest fights that we had with the internal teams. There were strategic fights as well. I think that whether or not BI should be owned within its own product family versus being integrated into Excel was another big thing we had to deal with. But at the time, D3 was still relatively un-vetted. And we were saying we’re going to bet on this external, open-source library and ignore all of the pre-existing libraries and the ongoing investments that Excel and other teams were making in their own vis stack. That was a tough one.
It was a lot of experimenting and proving to other teams that we had something here. It wasn’t that we ever convinced them to be clear. It was more than with the data that we would bring to the table would always be just enough to get us to the next conversation. And if you do that enough, eventually, you ship the product. You’ve just shown the value. And that was something we had to kind of like take as a tax, to be honest, all the way through toward the end.
Steve Herrod: Can you just maybe talk a little bit more about some of the pros and cons of trying to lead this entrepreneurial effort within such a big platform?
Nick Caldwell: I’ll put the cons aside. The cons are dealing with the existing sort of players. But the pros are, if you can get enough support and momentum behind you, it’s like you got a jet ski that has the same firepower as a battleship… The fact that we were able to spin up from a very small 15-ish person team to 300 people within the space of like a year and a half. That’s extremely hard to do if you’re a smaller company. In a big company like Microsoft, those resources can be pooled together very, very quickly.
“From a tech and product integration standpoint, once we went through all the hoops of convincing people that we were onto something from a product perspective, and our tech choices made sense. Once we had proven out the thesis, we got enormous leverage.”
Power BI went from not having really any significant usage to millions of MAU in the course of about a year and a half. We have to acknowledge that a big chunk of that comes from being integrated into this larger ecosystem. The downside, though, is internal politics. If you build something new in a big company means you’re disrupting something old. And big companies with well-established product lines want predictable revenue. They don’t want new things coming into existence, which might cannibalize something that’s already working. And they view it as not necessarily the upside for building something new and innovative. They view it more as the risk of disrupting something that is already in place. And that tension existed even past the MVP. We were just constantly trying to convince people that there was enough space in the world, for example, for Excel to exist alongside Power BI.
Quentin Clark: You were at Reddit during their scale era, during this time when Reddit had to get to really, truly hyper-scale. What was it like as you walked in? What were the biggest things that you triaged in terms of the challenges you had to tackle?
Nick Caldwell: I spent the first couple of weeks just meeting with leaders who were very hungry to meet me and express their vision for the company. And I kind of kept an eye out for people who would make good management material, people who cared about other people, people who cared about deadlines, and people who could be thoughtful in terms of technology, choices, and overall strategy.
“And I tried to explain to them that management didn’t mean sitting around and being told what to do. It meant a better understanding for everyone about the impact of what their day to day work would do. And predictability that would allow our product team and our other go-to-market functions to trust us more.”
And ultimately, these were the things that would be necessary to scale. And I think the other thing is I also identified were people who just didn’t agree with any of that. But for us to go from 35 people up to 100 plus, we’re going to have some semblance of structure. And I’m going to make sure we hire the best managers we can to take care of our people and be able to ship with some level of predictability.
In enterprise, it is possible and required, I would say, to build very coherent, clear roadmaps that can be communicated with the field as well as, to some extent, customers, so they understand what’s coming day-to-day. With consumer products, though, you have to think through… There will always be a layer in consumer products where you actually don’t know what’s going to happen. And companies tackle this layer in lots of different ways to try and rein in the uncertainty.
Big companies like Google or Facebook, they’ve got so much data that when you talk to those types of folks, they’re using data to rein in the chaos. Lots and lots of experimentation and so forth. That doesn’t work at smaller consumer products. You don’t have the data. So what do you do? You might see really small companies has the dial all the way turned up to customer interviews. Going and doing tons and tons of interviews and trying to get data in some other fashion.
“But the thing is, no matter which of these choices or methods you use, the consumer market itself is still just unpredictable. So you can do all the experimentation and all the different trials you want, but you have to internalize as an engineer or a product person that a lot of these things just aren’t going to work.”
And you have to be willing to pivot frequently until you find something that does work, gets traction, and then you follow that growth trajectory. And that that’s the essence of the difference… So the challenge that we had at Looker was that transition from sort of freeform execution with no roadmap and very little coordination between R&D and the go-to-market function to something that would be much more predictable and strategic, and build relationships between these functions that are required to work together as the company scales… You have to always find a way to help engineers understand what benefits they will get if they subscribe to more prescriptive processes.
“… when we started to talk about roadmaps, execution immediately got better because people understood what they were supposed to do and when. And the way I like to do roadmaps is to make it so that in the near term, we can do things very predictably, with very high accuracy, and people started hitting their goals.”
Equivalent to Magic
So Steve and I titled the podcast Equivalent to Magic to honor those moments when the tech really just makes a magical experience or outcome magically just work. So we always ask our guests this one question, what’s a magic moment in your career when technology was just magic?
I think for me, it was sort of a transformative moment. I know we didn’t cover this, but early in my career, I fell in love with the internet. The early internet, I guess bulletin board systems, and so on and so forth. And I decided to set up my own BBS. I don’t think you could do this, nowadays, by the way – 10-year-old set up their own BBS. I don’t think they would even allow that. But, well, for whatever reason, my parents allowed me to set up a BBS on our home phone line.
And I remember, one day, someone called in internationally from Quebec, and I had a conversation with this person. I grew up in Prince George’s (PG) County, Maryland, like 95% black suburb. I didn’t travel internationally, probably for another 15 years. Yet, here I was in the comfort of my home having an online chat with someone from Quebec in half French. And for me, that was magical. Like, it was a moment where I always loved tech. I always love coding. But it occurred to me that this was something that could give me access to opportunities and people anywhere on the planet. And that I didn’t have to limit my ambitions to PG County, Maryland. That with technology, I would be able to do, to meet or go anywhere. And for me, it was transformative personally and really changed the way that I thought about my career and my opportunities.