Atlassian Corporation (TEAM) Earnings Call Transcript

Oh, okay. Great, wonderful. Look, so we're just delighted to be kicking off day 2 of the conference with Atlassian. Seating to my right is Sri Viswanath, who is the Chief Technology Officer. And Sri, what we're going to do is we're going to go through your background a little bit, so people can get to know you. And then we're going to talk a little bit about how you manage 2,000 engineers with Atlassian.

Excellent.

That's got to be some trick. Yes.

I'm excited to be here.

And then we'll talk a little bit about the normal questions you'd get at a conference like this. But it's great to have you here.

All right. So first of all, so people can just know a little bit about you, where you're from, where were you born?

Yes. I grew up in India. I did my undergrad in India, came to U.S., did my master's in computer science.

And so what year are we talking about when you...

'96-'97. And then I did my master's, then I joined Sun Microsystems. I spent about 9.5 years at Sun, great company, learned lots of cool...

Those were the heydays, right?

Heydays. There were some heydays, then there were lots of things. But Sun was a pioneer in technology and also in people and caring for people. Every time I meet people from Sun, people are like, "Wow, that was a great time."

Yes. So that was, what, '90...

'99 to 2008 when I stayed there. So I learned lots of basics on people management and working on leading technology and pushing the boundaries of technology at Sun.

Do you have any thoughts on the -- and I totally did not prep you for this question.

No.

But if you don't want to answer, fine. But it just occurred to me, do you have any thoughts on the Oracle-Google lawsuit and whether APIs and the structure of Java is something that should be copyrightable?

It's funny. So I managed the J2EE enterprise server...

Did you? Okay. So you're the perfect person to ask.

We worked on open servers and we worked on -- we were pretty open when we were at Sun. It's like we didn't want to patent, we didn't want to -- we want it to be open. Opens have grown up thinking it's actually good for the society to have open APIs and not having patents for the interfaces so that there could be innovation in the ecosystem. All right, I'll leave it there.

No, leave it there. But it's fair enough, fair enough. And when did Oracle buy -- were you still there or...

2010. No, I'd left by then. I'd left by then.

Okay. So 2008, where do you go?

I went to Ning, which was Marc Andreessen's third company. And we were doing social networks, consumer space. We were building a platform. We were actually scaling faster than Facebook and we were bigger than Facebook.

This was the build-your-own Facebook, right?

Build-your-own social network. Yes, it's fascinating times. And we went through -- I was there for 4 years, and it was being in 4 -- 3 different -- 3 or 4 different companies, we were scaling fast. We did lots of scaling in the product and building the product for social network. Then we had to change the business model, so we went paid-only and then we had to lay off the company, too. We got profitable. We sold the company after that, so there's the sale portion. And then we landed at Glam and then we had to transition, so there's like lots of exciting...

But what was the lesson from Ning? And we'll do Glam after.

Yes. The thing that I learned at Ning was the leadership on how do you scale the platform. We were doing cloud-native before cloud-native existed, right?

That's 2008, right? Yes.

2008. There was AWS, but we couldn't use AWS as is. So we had to build all the services back in the day. We were pretty cutting-edge. And the skills that I learned and the leadership that I learned on you build it, you run it back in the day, when it's becoming a buzzword now, it's fascinating to see it. It's like, "Wow, we did that then. I should just tell people we did that then."

Yes. All right. And then so it ended up in Glam. And Glam was super hot for a while, too. And then...

Correct. Yes. So the reason Glam bought us was because there was -- they could do ads on social networks and they could grow. I was only there for a few months and then I transitioned out. And then I was at VMware for a short period, for a year, where I ran the Horizon Suite at VMware. And then Groupon...

Yes. Talking about that because that's interesting. So Groupon was what year?

Groupon was 2013-2016.

Yes. And how did you -- who called you? How did you end up at Groupon?

I didn't know people at Groupon. So they recruited me there. It felt like challenges, chaotic in terms of Groupon had acquired lots of companies and there was lots of platforms. And 2 things there. I had to rebuild the cloud platform and aggregate everything into a single global platform and the product across Groupon. And the second thing is we built a machine learning system that replaced humans in between. So that was fascinating because building machine learning has actually 2 problems. One is the technology itself, how do you build it...

Wait. What was it -- what were you trying to do with the machine learning?

So if you open up Groupon app, you see the deals. And you need to personalize the deals. And the top 3 or 4 deals is a significant percentage of revenue. So making sure that we understand what you like, where you've been and getting all the signals right. And it's actually a very complicated problem because it's not like you stack rank things, "This sushi place is better than that sushi place." It's not that simple. And we had city managers who'd stack rank things, "I know this sushi place, it's the best. That needs to be on top." But if you're 30 miles away from it in that city, you don't want to go drive 30 miles. You would just go nearby. So the propensity of travel matters a lot in different categories.

Yes. We had a -- and you never know what the -- some of the signals are going to be, right? So for example, yesterday, we had this company Eightfold. It's a private company, right? You know them?

Yes.

So -- and he was talking about the algorithms he's building to match candidates with jobs, right? And I go, "Give us an example of one thing you learned." And one thing he learned is there's a code when you're in the military, when you've had a certain kind of training. And that most of the people who get that training are the army chaplains, right? And those people really excel in marketing, right? Like, "This is fabulous," like, "How did you get that?" So yes, I love it. How long were you at Groupon?

About 3 years.

3 years.

Yes.

And so that takes us to, what, '16? Is that when you joined...

2016. January 2016 is when I joined Atlassian.

Okay. And then who called you? Was that a recruiter? Or how did that -- how did Atlassian happen to...

I knew a Board member at Atlassian. And he introduced me to Mike and Scott. And he's like, "It's a good company." I knew Atlassian for a long time. I had used Atlassian at Ning, at Sun, in fact, the early days, and also at Groupon. So every company that I was in, I was intimately familiar with...

Were you just using Jira? Or were you using...

Jira and Confluence.

And Confluence.

Yes. Both in all 3 places. And I had seen the power of how it can transform. And one of the things is for all those 3 places, I'd always use on-prem Atlassian products. And Mike and Scott were like, "We need to do cloud. We need to be a cloud company."

In '16?

In 2016.

Oh, that's interesting.

I was talking to them late 2015. So they're like, "We need to be a cloud company, so come help build the cloud platform," which is what I did.

Okay. Did you go to Australia to meet them?

No. They come here quite often. But I did go to Australia quite a lot the first year. I go there quite often. It's a direct flight, great city.

Yes. So what was the situation when you landed Atlassian in 2016? I mean when you looked around and you're like, "Okay, here's what I've got to deal with." What were the first couple of things you saw...

I knew Atlassian was really good, going in. When I went in there, I knew the products are good. There's lots of work to be done. When I went in there, I said, "Wow, there's nice people." I didn't know about the people itself. I didn't know anybody at Atlassian working. But Atlassian has such a nice culture, where people don't see from outside. It's actually really, really solid. I think that solid foundation has helped us to build different dev centers, different products and cloud-native and build a platform that cuts across all these products. These are hard people problems, right? The technology is actually straightforward. Like if you put smart people in, you can figure out a method to solve it. But the people problems are actually much, much harder. And having that solid foundation of having great culture, solid people like that, I think, has helped us grow a lot. So the surprising part was I didn't know how nice the people were inside Atlassian.

That's really interesting. So what's -- I don't think we think about that too often that the people problem is actually harder than the technology problem. What's the hardest part about the people problem?

Yes. The thing about people is when you have lots of teams and when you're -- and it's increasing in dimensions when you're moving to cloud because what's happening -- let me step back and talk about the industry, right? So industry is moving to cloud, like all of you know it. The things that happen behind the scenes when you move to cloud, you start using all these third-party products, like you use AWS, you use -- and you're moving to cloud and you're moving to micro services. When you move to micro services, you get these small teams of like 6 or 8 people, 10 people owning certain things. And then it becomes lot more -- lots and lots of small teams. In fact, we have like thousand services running in production. All these...

Each service has a team or a couple?

A bunch of services have a team. But lots and lots of small teams that are independent, have clear APIs, they have a mission, they run as fast as you can. So the problem becomes when you move to cloud and move faster, we do like 50 deployments a day, right? You need 2 things to do this, right? One is you need heavy tooling and automation across the board, right? You need to keep all of them aligned, otherwise they all go off in different directions and you don't get the output. So when you have all these teams, you build it, you run it. They want to be owning things as much as possible. Like you want to let them loose, but you want to let them loose on a specific direction, so tooling, extremely important. And number two, alignment and people aspect of they still need to work with all the other teams and making sure that they actually work well with different teams. And sometimes, you don't want duplication across all these things. So you need to start pulling common pieces and create a platform. Every multiproduct company will have to do this. And that's not an easy problem to take something away from a team saying, "I know you built this. They're also building this. I want to move it to a common place."

Give me an example of that.

Let's say if somebody is building identity. And you want to have a central identity service. If they are building attachments and uploads and media-specific features, you don't want to build it 7 times in the product. You want to move it to a common service. You don't want to build even -- there's different levels of things in the platform. There's infrastructure layers. You want to use AWS in the same way. You don't want to use AWS for something and GCP for something else. And you want to have a strategy that works across all those. So there are structures that you need to put, and you need to have these common things that everybody needs to follow. But you need to make sure that they're self-contained units that they can go as fast as possible. So people problems are actually extremely important. And the beauty of Atlassian is because we have an awesome, solid foundation in terms of culture, we actually try to build in into our tools. We dogfood every single product, as you can imagine, across our Jira and Confluence and Trello and Jira Service Desk. We have lots and lots of Jira services. Every single team, like IR team, for example, uses Jira Service Desk for interfacing with the request coming in from different teams on all the things that we do. I'm sure there's a Jira service ticket for JMP conference to figure out what we need to do and who needs to engage and track all the things...

That's why you guys showed up early in your own time.

There you go. Look at that. We're super organized. But the beauty is for all our developments -- and development has gone beyond just engineers, right, so it touches every single part of the organization. And every company is becoming a software company. So our products are actually used by every single one -- even though you start with, "Oh, it's a technical product." It's actually not a pure technical product because we have program managers, product managers, IR people with finance people [ FDN ], everybody uses -- touches our products. And we need to make sure that it works for engineers but works for everybody else. So we dogfood, we use our products internally ourselves across the board.

Yes. Well, it is -- I mean that was the point, right, which is that the biggest problem you have is teams. And of course, the problem you're solving with teams. So you feel great, right? So that should work.