Microsoft’s recent acquisition of GitHub has spurred a lot of conversations regarding which company is the next big acquisition target, and here on DevOps Chat is no exception. In this episode of DevOps Chat, I speak with Rob Zuber, CTO of CircleCI about the CI/CD space, the future of the market and, yes, M&A activity in the DevOps space.
As usual, the streaming audio is immediately below, followed by the transcript of our conversation.
Alan Shimel: Hey, everyone, it’s Alan Shimel, devopsy.kinsta.cloud, and you’re listening to another DevOps Chat. Today’s DevOps Chat features Rob Zuber, CTO, CircleCI. Rob, welcome to DevOps Chat.
Rob Zuber: Thanks. Good morning.
Shimel: Good morning.
Zuber: It’s a pleasure to be here.
Shimel: Or afternoon, in my case, or evening. Who knows when people are listening to this? But, Rob, thanks for being our guest. I think this is the first time we’ve had anyone from CircleCI on a DevOps Chat here. So, you know, you are a mold-breaker and thanks very much.
Zuber: I’m excited.
Shimel: All right. So, Rob, our audience tends to be pretty DevOps-y, and so I’m assuming they know of CircleCI or at least have heard of CircleCI, but maybe they haven’t. So, in case they haven’t, Rob, give us a little background, just a real quick elevator, on CircleCI.
Zuber: Yeah. So we’re a big part of the delivery pipeline for many software organizations. We offer continuous integration and continuous deployment, both as a cloud offering so that you don’t have to manage any of it yourself, but, if you’re interested in doing that, we also have a server-hosted version. And, you know, our objective is to help people focus on their business and get their software into market faster, taking out of the way the overhead of managing a lot of the tooling around doing that.
Shimel: Yep. And, Rob, just in way of history and background and setting the table here, CircleCI really is – it was kind of a cloud-native CI/CD solution originally and it’s only relatively recently that you started having sort of an on-prem solution as well, offered as an option. So it’s kind of opposite of the usual SaaS migration, if you will, right, where people turn on-prem stuff into SaaS solutions. You were kind of a SssS solution and now offer an on-prem option.
Zuber: Yeah, that’s exactly right. And that’s probably partly a function of when the company was started, so CircleCI launched in 2012. At that point, people were pretty comfortable with cloud offerings. GitHub had been around for a little while, so people were used to having the source code in someone else’s servers, and so we were able to leverage that a little bit. And we were able to grow quite effectively off of that base, but we did reach a point where there were very large customers with very specific needs, where it made a little more sense for them to run the software themselves and be able to have a little bit more opportunity to customize the environment in which it was being run, etc.
One thing that’s interesting about that – and you talk about the difference in which way you go across that transition – because we built a multi-tenant platform from the beginning and really focused on our ability to operate that on behalf of other people, we find that, even in an on-prem solution, especially for very large organizations, we have a lot of the capability around that management at scale and centralized management that doesn’t – that still gives freedom to the individual developers to do their work. So it is actually quite interesting, how the genesis of the product affects how it operates in those multiple environments long-term.
Shimel: Absolutely. Rob, another area – and I didn’t really talk to you off mic about this, but another area I wanted to explore was, historically, was CircleCI originally just CI or was it always CI/CD?
Zuber: I think you could kind of look at that two ways. I mean, there was CD pretty early in the product, at least for some particular cases, but, if you look at the evolution of the software industry, many more people were doing CI in 2012 than CD. I mean, it was sort of this slow transition of people getting to, you know, “First, we’re gonna figure out Agile process and just how we do our actual software development. And then we’re gonna add CI on top of that. And then we’re gonna add CD.”
So we always had some CD capabilities, but the number of our customers or the percentage or breakdown of our customers using CD has grown pretty significantly over that time, and I think that’s partly an effect of the kinds of tooling we’re offering, partly a shift in what CD even means, and then partly just growth in the market and comfort level of people doing that, right? Of – especially – and CD, of course, is always challenging ’cause some people are talking about continuous delivery and some about continuous deployment, but, in particular, on the continuous deployment side, other things had to come into play – you know, feature flagging, canaries, just better control of what it meant to actually put something in production. And so the comfort level and increase in that has led to more people using our platform for CD as well.
Shimel: Absolutely. You know, as an observer of the space, sort of the market, I’d seen it come – when I first launched devopsy.kinsta.cloud, four years ago, Rob, there was this clear sort of distinction between “We’re not CD. CD is something else. We’re CI,” or reversed – “Those folks are just CI. They’re not CD.” And it kinda just seemed like an artificial wall had been built, right? Not to get all political.
But – and so I, for one, am happy to see sort of that wall being torn down and the natural flow of the software development life cycle, of the modern software factory, if you will, taking place, where CI kinda naturally flows into CD. And I think the next thing we’re seeing is this natural connection to what Gartner and some of the others call “ARA.” Right? Which I don’t even know what the difference between that and CD is, but I don’t know. What do you think?
Zuber: Well, I think that, overall, this evolution, as I was saying, it always takes a long time for people to get – or, let’s say, for everyone to get comfortable with each of these stages of the evolution. And we’re always looking to the next frontier, right? So, at one end of the spectrum, you have those early adopters, who, in 2012, were doing CD or before that. I’m trying to think of the original sort of posts around CD that I think came out of, I wanna say, Etsy, but I don’t know if that’s actually true or me projecting kind of them as a pillar of this type of thinking. But, you know – and we’re still at a point where some people are just getting comfortable with and adopting it, so the spectrum is really broad, from early adopters to sort of overall mass adoption.
And the early adopters have moved on to, probably, things that I haven’t even thought about, and so we’re always seeing that and it’s nice to have people try those things and, in some cases, fail, and, in some cases, discover really new and interesting ways of approaching stuff. And it takes – it definitely takes a little bit of a mindset of “I’m just willing to try this. I’m willing to take some risk and do new things,” and the people that do those create opportunities for the rest of us to learn from, the things that go well and the things that go horribly wrong.
And so it’s one of the reasons that I just love being in the space in general, is just the constant kind of desire to do things better and find new opportunities, of course, with a good balance towards actually focusing on getting stuff done, but there’s just such a creativity and desire around improving, in any way we can, this process and how we get stuff out and how we build it reliably and work efficiently.
Shimel: Absolutely. So, Rob, let’s turn a little bit to – we’ve done a nice history. Let’s look a little bit to the future now, though.
Zuber: Mm-hmm.
Shimel: Talk to me – what did – how – you’re the CTO. You kinda have a lot of the vision for CircleCI. Where do you see CircleCI and the market in general going?
Zuber: Yeah. Well, that’s a great question. I think a lot of people’s perspectives of this market was changed a little bit in the last week or so, with news around Microsoft and GitHub and just the value that people are starting to recognize in developers, the happiness of developers, the efficiency, the value that they place on the tools that they use. So, for the market overall, I think we’re just a really exciting time, where we talked about this – or I talked about this creativity and passion to do things better, but we’re at a sort of inflection point where there’s a really, really high-level recognition of the value of that.
I would say there’s been a significant shift in the mentality of many companies. And we talk a lot about how every company is becoming a software company, in that, even if you’re an airline, your biggest issue, when you ground your planes for a day, is that you have some software glitch somewhere, right? You know, multinational trading markets are shutting down on software issues. So the impact and value of software is becoming really, really visible and, as a result, there’s a shift in mentality about how we build software, from “This is a cost center that I just have to manage down,” to, “This is the core of my business and I need to be really investing and doing this in the best possible way.”
And so that’s obviously exciting for us. And one of the things that’s really interesting for us, as we look forward, is that we – we’ve had the fortune of some success and growing and, to the best of our knowledge, run the largest or one of the largest build platforms on the planet. As a result, we know a lot about how people are building software, and so we’re investing a lot of our time and energy in figuring out how to use that to then help those people, meaning, if you’re building on CircleCI, you don’t just get an understanding of your particular code change that you just made, but you can get an understanding of how your team is doing, how your whole process is doing, how effective you are, relative to other teams out there.
And I think that’s a place where we have a lot of opportunity to really add additional value for our customers, in not just, again, direct feedback on this one thing that you did today, but really helping understand your process, maybe where things aren’t working well, maybe where you have some opportunities to improve. We’re all in this, trying to be better every day. And if we can help customers at that at a higher level, that’s really exciting for us.
Shimel: Yep. You know, it’s interesting you brought up the GitHub thing and to justify all this. And so I have some views on it. What a surprise. But, you know, Rob, couple things. No. 1, when you look at GitHub and the price Microsoft paid, it obviously is a ridiculous multiple to revenue. I saw a chart earlier today. I forgot if they said it was 21 or 24 times revenue, which is wow. Right? That’s why people do software companies.
But another way of looking at it is, look, Microsoft has always coveted the developer community. And, over the last – I mean, for as long as I’m in technology, 25, 30 years, they’ve done not a crappy job of working with the developer community. Yeah, there are haters and Linux folks and the open source, when Ballmer was there, you know, but they’ve always gone after the developer community hard. And so they had an opportunity to buy a company that has, some estimates say, 28 million people or potential developers, right, in that community that GitHub has. So, when you look at 28 million and what they paid – don’t get me wrong; it’s still a lot of money, but it’s a little bit – you could see where they saw the value, okay?
Zuber: Mm-hmm.
Shimel: No. 2, having been in the tech, software infrastructure business a long time, I learned a lesson from my friend Brad Feld, who I sold my first company to and he financed or was one of the key VCs behind many of the companies I’ve been involved in, and that is, look, first mover or first – not the first mover, first acquisition in the space always gets the lion’s share and the best multiple. The second one does okay. The third one does a little worse. If you’re not in the top three, don’t bother. Right? That was something Brad always preached to us.
So, when you look at Git – or GitHub, excuse me, being the first kind of acquisition in that space, getting this kind of multiple, they deserve it. Right? They were first. They were the first ones bought; they’re gonna get the lion’s share. I think the message to other folks in the GitHub space, specifically, is “Hurry up and get bought ’cause you don’t wanna be number four or five.” Right? You don’t wanna be left in the musical chairs game when the music stops.
Part and parcel of that, though, Rob, is that companies like Microsoft, like IBM – we used to say “like HP, Cisco,” you know, big, big public entities – it’s very hard for them to be innovative. It’s very hard for them to do their own R&D. It’s very hard for them to build these communities themselves. And so what they would spend in R&D or community actually goes into their M&A budget and they buy those things.
Zuber: Mm-hmm.
Shimel: So, when we look at a GitHub, we look at a CircleCI, we look at any – a Chef, a Puppet, any of the – even the Jenkins or CloudBees, any of the players in this DevOps space that is so hot right now – GitLab, another one – they all are gonna be very attractive to companies of a Microsoft kind of girth. Right? Because the big guys can’t – they can’t do what you’re doing, Rob. They don’t –
Zuber: Yeah.
Shimel: You know.
Zuber: Yeah. I mean, I think this is classic innovator’s dilemma, right?
Shimel: Yeah.
Zuber: Did I get that right? So it’s difficult. I mean, if you look at the first five year – I mean, we all talk about these overnight successes, right, and they always – they’ve been around for 10 years. And the first five years many of those companies is extremely painful, it’s a grind, and the incremental increase, the amount of revenue you’re driving or whatever, would be so insignificant on any one of these companies’ radar that they just – they won’t put the time and energy into it, right?
Whereas you’ve got a company like ours or anybody in the space, honestly – I’ve done a bunch of different startups in different spaces – you’re living and breathing it, right? And you’re so tuned into what’s happening that you can see it. And that kind of incremental growth, to you, is everything and it’s super exciting and you get this passion behind it that allows you to build a very focused and, honestly, a great product. And I think GitHub is a great product. And then, suddenly, you have 28 million users – I’m quoting your number so I’ll assume you’re right –
Shimel: _____ I’ve seen.
Zuber: And a massive community. You know, and then, going back, as you were talking about the time and tech, I mean, I got into tech with the quarterly mailing of MSDN CDs and, honestly –
Shimel: Exactly. Yeah.
Zuber: – my favorite IDE – I’m a Mac and Linux user, almost exclusively, but I used Visual J++ in the late ’90s, when I started doing Java development. It’s still the best IDE I ever used. So Microsoft has been good, very, very good, with developers, but the kind of developer community has started to drift away from where they were going. And so this is – I wouldn’t call it a course correction because they still have all of their community, from a .NET and sort of Windows world, but this is an opportunity to be less about Windows, you know? And we’ve seen Microsoft shifting away from –
Shimel: Oh, yeah. Satya Nadella’s Microsoft, Rob –
Zuber: – just pure kinda _____ –
Shimel: Not to step on you, but Nadella’s Microsoft is a very, very different place than when Steve Ballmer was there. Right?
Zuber: Yes.
Shimel: And they are moving to Azure and though Azure does Windows well, it does containers well, it does Linux, it does – you know, there’s very little it doesn’t do.
Zuber: Right.
Shimel: You know? And I will tell ya – I’ve told people this – I think, one day, Microsoft buys Docker. That’s my personal opinion. I think it’s another kind of – not to get into it or badmouth or anything like that, I just think someone’s gonna buy Docker ’cause the revenue has to justify the valuation, and so – and Microsoft’s a great candidate to do so. But, anyway, you heard it here first. But, Rob, that’s all fine and dandy. GitHub plays in sort of a different segment of this DevOps market than you guys do, certainly.
Zuber: Mm-hmm.
Shimel: And we really haven’t seen the consolidation yet, that I think we will in the CI/CD and, if you wanna call it, ARA space, but it’s coming. I’m sure.
Zuber: Yeah, we’ve seen movement in the market, but most of it has been off the bottom end, if you will, meaning sort of the smaller people we would consider to be our competitors but didn’t quite reach critical mass and then got pulled in by other companies, for the people or because the product was going to be useful to them internally, that sort of thing. Not a roll up and now we’re taking this and getting it scales or _____ –
Shimel: So we used to say, when I worked with Brad, those are companies that really were a feature, not a product.
Zuber: Mm-hmm.
Shimel: But, with that being said, Rob, we went off on a little bit of tangent here. We’ve used up way more time than we were supposed to. But maybe we could have you back on and – ’cause I wanted to really get into a little nuts and bolts around continuous integration, delivery, what it means for developers, what it means for ops folks, but why don’t we do this? We’ll call it a day on this 25-minute DevOps Chat and we’ll have you back on maybe next month. And let’s get into what it means for developers and ops folks in even testing, in this new world of CI/CD.
Zuber: That sounds great. I’m always happy to talk about that as well. There’s so much happening in this industry, in this space right now, that it’s hard to keep it short. [Chuckles]
Shimel: You know what? Hey, man, it keeps my kids going to college, Rob. [Laughs] Anyway, hey, Rob, Zuber, CTO, CircleCI, this episode’s guest on DevOps Chat. Thanks for joining us. This is Alan Shimel. You’ve just listened to another DevOps Chat. Have a great day, everyone.
— Alan Shimel
By providing deep hooks into the Azure platform, Lam says Talend is looking to make it easier for teams to build applications by relying on an integration platform-as-a-service (iPaaS) environment to access data stored in Microsoft cloud services such as Azure SQL Data Warehouse, Azure Databricks, Azure Data Lake Store and Azure HDInsight.
