February 1, 2018
LaunchDarkly: (How To) Test In Production
Heavybit member company LaunchDarkly held their Test in Production Meetup in the Heavybit Clubhouse on January 10th. In this session the foc...
In episode 46 of To Be Continuous, Paul and Edith discuss Microsoft’s acquisition of GitHub. Is it a good thing? What will be the impact of such a massive deal on the companies and developers in the devtools space?
Edith Harbaugh: Hey, Paul. Microsoft just announced that they are acquiring GitHub.
Paul Biggar: I already tweeted something about this yesterday.
Edith: Way to date the podcast.
Paul: I know, yeah. How they can figure it out.
Paul: Also, they did it yesterday.
Edith: The tweet storm was pretty good. How did you come up with it?
Paul: I think it's just the way I come up with all of my tweets. I look at the comments on Hacker News, and I've closed the thread angrily at how stupid all the comments are. And then whatever is still in my head 15 minutes later might get tweeted.
Edith: Well, all right. So anybody can go look up your tweets, but do you just want to summarize your thoughts around the acquisition?
Paul: I think in general it's very good for our entire space. A lot of people are anti-Microsoft, and I'm not. I think Microsoft is on the right path and is financially incentivized to be to be good for open source. But also great exit for devtools, which people said there was no money in. Great exit for a lot of people at GitHub. There's going to be a huge crop of agile investors in the devtools space in the next couple of years. Just great for everyone.
Edith: I thought your tweet storm was great, but I'll summarize a couple of things that jumped up for me. Just how much bigger it was than other devtool acquisitions.
Paul: Yes. 7.5 billion--
Edith: Billion with a "B."
Paul: Yeah, "Billion with a 'B.'" And a couple of years ago I saw someone tweeting about, "Maybe Atlassian should buy GitHub." And Atlassian couldn't buy GitHub. Couldn't buy the value.
Edith: I have some gossip which I will not attribute to anybody. For just some context, GitHub's 7.5 billion. The next closest acquisition in the devtools space had been Heroku for 500 million.
Paul: Heroku's 250.
Edith: 250 million. You're right. And Xamarin for a half million.
Paul: Xamarin was around 500. There's another one--
Paul: Is Trello in the dev space?
Edith: They got acquired by Atlassian. I mean, I would put that in the dev space.
Paul: What was that API company that Google sort of reverse-acquired? Where it went public, and then Google acquired it?
Paul: Apigee, yeah. So that was a billion or 600 million, or something like that? It wasn't small.
Edith: They'd already gone public though, and then they got acquired. And MuleSoft just got acquired too.
Paul: There's an interesting thing, that those are classically in the enterprise space. And the other companies that we were talking about, like Xamarin and so on, those are like the bottom-up devtools companies that I think of as more of the of the devtools space.
Edith: That's interesting. Because I've heard that most of GitHub's revenue came in-- I mean, it's publicly available. It was all coming from GitHub Enterprise.
Paul: Well yes, and that's not what I mean exactly. Because I think everyone in the devtools space makes like, "Live long enough and you will make all of your money from enterprise."
Edith: "Or die, because you won't go--"
Paul: "Or die." Either because you don't know, or--
Edith: Which has happened to some friends. You know who I'm thinking of.
Paul: I know many who this fits.
Edith: Many, many people who have gone through Heavybit and not turned to the enterprise.
Paul: I wasn't thinking to Heavybit, but just in the dev space in particular. You get an e-mail from people every now and then saying, "We're closing down. Do you want to buy us?" And it's like, "Oh. I could've--"
Edith: I think part of it is, and this is a little bit of a detour. But we'll come back.
I think some people just want to run a foundation, not a startup.
Like I was using this example of somebody that does, and I don't want to say exactly who they are because I think they're nice people. But they are doing a large registry project and they just don't seem to have the fire in the belly to make it to a billion dollar business. I think they'd be quite happy if they had two million dollars a year and were running a foundation.
Paul: I think this is one of the problems with startups being so prominent in the software industry in general, because it becomes like the cool thing to do. It becomes the way to fund your thing, it becomes the way to make money. There's a lot of people who just don't want to do that. They want to make a bootstrap business, they want to make a foundation, especially around the open source world. There's a lot of people who want to find a good way to fund open source. We talked to Nadia about that before.
Edith: Once you get on the VC train, you're on a train.
Paul: Yeah, and it is a struggle to get off that train. And when you do it usually ends badly. Or ends perhaps not badly, but not well.
Edith: I mean, just to walk through it. It's considered a huge exit for GitHub and for Andreesen. Because they put in 100 million which is now--
Paul: A billion-ish, I think.
Edith: Which seemed ludicrous, and now gave them a 7x return.
Paul: I remember Tom Preston Warner talked at startup school, this was 2012 I think, maybe just after they had raised that. And in the same thing, Spolsky was talking. And one of the points he made is that we don't know how many developers there are. I think he was at the time estimating there's like 20 or 25 million.
Edith: Not just 25 developers--?
Paul: In the world. Right. But a few years beforehand our best estimates were that there were eight million developers. Stack overflow showed us, "How could there be this many users?" And what we're seeing, GitHub had 29 million listed users. And not all developers are on GitHub. In fact, I would say a majority of developers are not on GitHub.
Edith: I think it really shows the power of developers. I think it is very good for startups that are eager to go down that path.
Paul: The VC path.
Paul: I think it's good for both. I think the fact that there is big business to be made for devtools is good even if you don't want to go down that path. Even if you want to have a small SaaS business. I think the challenge with a SaaS business is things change a lot, and that can be a struggle for a SaaS business. But other than that I think it's good for VC backed businesses as well as bootstrapped.
Edith: So your first company Circle got funded originally in--
Paul: 2011 was our first funding, and that was like 50k. Then it was about a year later we did a million and a half, and then a year later we did another six.
Edith: What was it like to try to go out and raise for a devtool in 2011?
Paul: There were believers.
Edith: There were believers?
Paul: The believers mostly had been involved with Heroku. It was mostly in the Heavybit space. So Michael Dearing and Steve Anderson were the big checks in our thing. SV Angel wrote a big check, or a medium check I guess. Big for them. And there was a lot of angels there who themselves were developers, or had believed in the thing. And they're making money on the thing, Circle goes well.
I think that without the first generation of tools, which I think of as Heroku and GitHub. Having shown a lot of success, and at that point Heroku had sold and GitHub had raised a billion-ish. But I think Stripe was not the crazy whirlwind it became, nor had the New Relic exit. New Relic at the time was not regarded as the success that it necessarily became. What's Twilio's market cap? Or 5.5 billion?
Paul: Yeah, and those had not happened.
Edith: My joke is that Heavybit is the house that Heroku built.
Paul: Oh yeah. Very much so. Which is where we record from.
Edith: I mean, I personally owe a lot to Heavybit.
Paul: Oh, same.
Edith: That's how we got the podcast. That's how I met you. And it's all because Heroku. So just think about what sort of things are going to happen now with the GitHub liquidity.
Paul: Oh absolutely. And the GitHub thing was 30 times bigger than the Heroku exit.
Edith: Just massive, massive, massive return for everybody.
Paul: For everyone involved. The three founders are probably about a billion each.
Edith: Wow. One of them, it's funny. Tom went to Mudd with John Kodumal, my co-founder and me, and I never met him. It's funny how it's a small school but a lot of comp-sci people came out of there. Joe Beda also went there.
Paul: And it's interesting with GitHub, it was a bit of a mess on the way.
Edith: Oh my gosh. "Mess" is a very generous way to describe some of the stuff that happened there.
Paul: It really is. There was the CEO being fired, Tom Preston Warner. Chris stepped down about nine months ago. I'm forgetting some things. There must be more things than that.
Edith: Those are the two main scandals that I remember.
Paul: I guess the other one, that industry back chatter, is GitHub was fairly famously flat. In the way that Stripe was, and we copied them both at Circle and that was a huge mistake. I think that everyone who went down that route had a disaster along the way.
In GitHub's case I believe they had big layoffs a couple of years ago as they were straightening out the internal culture. But I think what it proves is one, they did a good job for some financial definition of success, in turning it around. They were able to hit the next inflection points of growth and all that sort of thing.
And I think the other thing is that no matter how much of a mess your company is, so long as you have this really amazing product it survives.
Edith: I read a really interesting quote. Part one, I have heard through back channels, that the reason why the price went so high is there was a bidding war.
Paul: Interesting. Do we know who the other sources are?
Edith: Totally gossip. But the suspects were Microsoft, Google and Atlassian.
Paul: Atlassian. Interesting.
Edith: And that the bidding started around 1.5 billion. Which was because their last round was around that, and they were just trying to basically get sold at this point. Because the quote I saw was that--
Paul: This is good gossip.
Edith: Oh my gosh, yes.
Paul: Yeah. Wow.
Edith: So what I read, it was just like, "They could not go public." Because they didn't have a CEO, they didn't have a strong growth story. So they put themselves up for sale. What really inflated the price was just--
Paul: Three buyers.
Edith: Yeah. And Atlassian just does not have deep enough pockets to do 7.5--
Paul: No, Atlassian can't put 7.5. Atlassian's value is 15.
Edith: Yeah. It came down to Microsoft and Google, and those are two of the most deep-pocketed people.
Paul: I mean, I don't know where you get off the bidding train but I think I could have seen it go higher. It would not have been completely shocking for it to go. Considering that both of those are investing in trying to make 100 billion dollar businesses.
It's not the devtools, but it's a space that is led by devtools. But yeah, 7.5 billion. It seemed on one hand expensive for what we know about GitHub's revenue, but for its place in the ecosystem at the same time it seems like a steal.
Edith: I don't have any insider knowledge. Just what I've read publicly is that the revenue is between 100 to 200 million.
Paul: Yes. They published it. Or there was an article that in the last nine months it was, I think, 98.
Edith: So just to get bought at a 75x multiple is crazy.
Paul: Although that was nine months, so it's like 130. I put it at 50 but, yeah. Same.
Edith: It just seems like it was one of those things where you have to very deep-pocketed suitors.
Paul: Heroku had a similar thing. I don't know if it was bid up, but it was certainly that Salesforce really wanted it. It had revenues of two million on 250. Or Sorry, it sold for 250 on revenues of two.
Edith: It just becomes this thing.
If you have to have something, what's the right price?
Paul: Right. 7.5 billion, is that a future of LaunchDarkly?
Edith: Thanks for wearing the shirt today.
Paul: No worries.
Edith: I appreciate it.
Edith: I'm still working through what it means for the industry, I think. I wonder if it might kick off an M&A boom.
Paul: In the space? I think there has been one underway for a while. This certainly isn't the first purchase of devtools by, in particular, by the big three. By Microsoft, and Amazon and Google.
Edith: I think it shows that there's more money at play.
Paul: Azure GCP and AWS. Each of them are going to be huge. I would say that they can each be 100 billion revenue businesses. And so when you're playing at that stage, what difference does it make? And the amount of money that they pay engineers, like a mid-level Googler is making 650k, let's say?
Edith: Oh my god.
Paul: I know.
Edith: I'm in the CEO thing. I'm in the wrong business.
Paul: We chose the path we chose.
Edith: I did, I don't regret it for a second.
Paul: And there were reasons that we chose those paths. But you can go to one of those and if they're paying a team that's not performing, or suppose they've got a team that's trying to do something and it's not succeeding. And they're putting 10 million a year at it. And it's like, "Looking over at your startup for 20. No big deal."
Edith: So my summary is I think it's really good for the ecosystem. I think it unlocks a lot of money that could get reinvested into new start ups. I think it's good for proving the power of developers.
Paul: I think we're probably beyond the points where people doubt this, and I think maybe GitHub is the thing that puts us beyond this. I know investors have, for the most part, not raised the shadow of "developers don't pay for tools."
Edith: They did with us.
Paul: Yeah, but it's the ones who don't really know.
Edith: I think it was also that we are a new category. And we hear this even when we go talk to some potential customers. They're like, "Why would I pay for that?".
Paul: Right. There's a free, or there's an open source thing out there.
Edith: Yeah. And then the smart customers, they literally show us their calculations. They are like, "It would cost us this many sprint's of this many developers to do that. We'd much rather just buy your product."
Paul: I've probably mentioned this before but I have an e-mail from Paul Graham saying, "Developers don't pay for devtools."
Edith: Are you going to hit reply today?
Paul: I think when Circle is involved in the transaction, then that'll be the day I might hit reply on it. I remember seeing Paul Graham tweet about GitLab. GitLab raised some 20-30 million B, or something like that. And this was the thing he responded to, and he was like, "For something of this magnitude in developer tools, something big must be happening."
And I'm like, "Yeah. Developer Tools is happening. And it has been happening for a while."
Edith: Yeah he didn't notice, but developer tools are here.
Edith: I think it's awesome. I think there's also a lot of interesting corollaries to their acquisition of LinkedIn. In the sense that it's the network that they're acquiring?
Paul: Right. The product is you could build the product. It is challenging to build the community. What do you think about all the people jumping ship?
Edith: Who jumping ship?
Paul: Oh, all the people on Hacker News are tweeting about moving off GitHub now that the "Big Beast" owns them. I can see your face, so I share your opinion.
Edith: What does my face say?
Paul: It's a frustration at the people who, I guess they're the sort of people who won't pay for software.
Edith: And also I think, and it may be biased, but there is a new Microsoft.
Paul: I completely agree.
Edith: We went up to build, we podcasted from there. It's not the Microsoft of 20 years ago.
Paul: Not at all. We're two CEOs since that. When did Ballmer resign? Early 2000s?
Edith: I don't remember. But the people that I know from Microsoft are so smart, and so passionate. You have Nat, you have Javier--
Paul: The new GitHub CEO.
Edith: Yeah. Well, let's come back to that. You have Javier Soltero, and Kevin who was one of our first guests who was the engineering lead at Acompli.
Paul: They're also hiring a lot of people. Especially in the Linux and the container space. Like Jess Frazelle, and a lot of well-known Linux-ers are working at Microsoft.
Edith: I think people have the image of Microsoft that is unfounded.
Paul: Let's be fair, it was accurate at a time.
Edith: It was accurate at the time, but I think the people I know at Microsoft care deeply about software development. Like Cindy Alvarez is such a brilliant, brilliant person.
Paul: I think the thing about Microsoft is that it is a big company led by growth. And they care about that growth. And in '99 Linux was a threat to that growth, and open source was a threat to that growth. And their response to it was unpleasant.
Edith: My saying is, "They tried to create a walled garden to keep all this evil out." And I think they realized that they had just basically walled everybody else out.
Paul: I think they did more than that. I think when you look at Bill Gates' comments on Linux, a lot of the FUD wasn't coined for Microsoft, but it was used a lot in the, I'm going to call it the "Slashdot era."
Edith: It's interesting. In full disclosure, LaunchDarkly, my company is a partner with Microsoft. And they've been a very good partner to us, they came down to Heavybit because they wanted to see what the devtools companies down here were up to. And they asked people like, "How many Microsoft products do you use? Are you usually in Visual Studio?" And people are like, "What? You have devtools?"
And they're like, "It's not even that people dislike our tools, it's that people don't even realize we exist anymore. We are irrelevant."
Paul: Yeah, that's right. And they were long irrelevant. And this is the way that people felt about Azure when it came out. It was like, "They're making a cloud for some .net thing, and no one cares about .net." And obviously Xamarin, amongst other things, has proven that there is a large contingent of people who care about .net. But not in the Linux world.
And so they're busy remaking their image around, frankly, where the money is. Because the money is in Linux. It's in Linux hosting and no one really cares. I mean, there's a lot of large enterprises out there that build on Windows. But it's not a big deal.
Edith: I think what they realized, and again I'll say Microsoft has been a wonderful partner to us. I think they realize that they basically lost a generation of developers.
Paul: Completely true. I think that that's extremely accurate. I think they look at the innovation that happens in--
Edith: Ruby, Python.
Paul: I think we can call it, "In the open source world." I think it's fair to say that there's the Microsoft world and the open source world, and those are the two major things. But in the Microsoft world, Microsoft has had to make the innovation.
In the open source world it comes from dozens and dozens of places. All of our startups are, it's not that we don't have .NET SDKs or that kind of thing, but we're definitely servicing the open source part of the world.
Edith: Then we had Martin Woodward talk on an episode with us, who was spearheading open source at Microsoft. And how he's just like, "We've got to start innovating again. We cannot afford to be--"
Paul: I mean, now they're the biggest contributors to software on GitHub.
Edith: It's funny. I'm having a total flashback. As I said before, LaunchDarkly is a Microsoft Partner. About a year and a half ago in November I went to a Microsoft Partner event in New York.
The two presenters that stuck out in my head were the two non Microsoft presenters. One was Chris from GitHub. Came on stage and it was this big reveal.
Paul: Yeah, he was the surprise wonderful guest.
Edith: And I just remember everybody was shocked. Because it was such a culture thing of San Francisco open source versus Seattle .net. And then they flashed up, "Microsoft loves open source." And I was there, and it was just this moment. And I also--
Paul: When was that?
Edith: It was November of 2016. And then I remember Nanette Friedman onstage. Nat Friedman always gives a good talk.
Paul: This is before Xamarin got acquired?
Edith: It was right after.
Paul: Right after, ok.
Edith: So he was on his big publicity tour. I think it's a really good move for him. He's a smart guy. It's a huge--
Paul: For not a CEO at GitHub.
Edith: It's a promotion. I don't know if it's promotion or lateral, or just a lot of pressure.
Paul: I think from Xamarin when he went to Microsoft, I think he's head of developer services. Is that right? VP of developer services, I want to say?
Edith: It's funny. John was meeting some Microsoft folks and I tried to describe people's different roles and hierarchies, and it was a total mess.
Paul: Yeah, those--
Edith: This interleave of program management and VPs.
Paul: It's safe to say he was a big deal at Microsoft already. It wasn't an acquisition where he went and sat in the corner.
Edith: The Wall Street Journal actually did an article about this, about how they had this new rising crop of talent from acquisitions. And the two people they specifically called out that I remember, were Nat and Javier.
Paul: Right. Because Javier took over, I want to say, all of Outlook? Or all of mobile, one of the two.
Edith: For the listeners at home, Javier Soltero was the co-founder of Acompli which was not a Heavybit company but sat here.
Paul: So they sold Acompli to Microsoft and became Outlook. The new Outlook mobile.
Edith: What was the price, like 300?
Paul: I thought it was 200.
Edith: Something quite high.
Paul: They were a year old, two if even.
Edith: But they just had, if you're interested, for those listening. You can go and listen to Kevin's podcast with us about their engineering culture, which was brilliant.
Paul: One of the first ones, I think.
Edith: He was our first guest after Chris Gill.
Paul: OK. So, Nat is CEO now, of GitHub. They've been searching for a CEO since Chris stepped down about nine months ago, and my understanding is that he hasn't been running it day to day. That someone else, maybe their COO, was running it day to day.
Nat has not taken over yet, but will take over when the transaction closes. Which does not look like it's going to be a quick transaction.
Edith: Well that's part of the scuttlebutt I heard is, that basically they couldn't go public.
Paul: Because they had no CEO.
Edith: And they couldn't go public. It would not have been impossible for them to raise more money, but it would not have been at awesome terms.
Paul: People aren't going to invest in that sort of climate.
But acquiring and that sort of climate is totally fine, because you're going to make a few changes anyway. And the fact that changes are needed makes it easier.
Edith: I think they sold at precisely the right time. They could get a bidding war going.
Paul: Yeah, absolutely.
Edith: Some people were speculating before this acquisition happened that they would have to get sold in a down-round for like a half billion.
Paul: Jesus Christ. I remember when the day before it was announced, so Sunday night. People were speculating.
Edith: Well, I sent you a link.
Edith: I was like, "We need to record." And you were like, "What?!".
Paul: And we didn't know numbers. And I don't think I'd have guessed 7.5. People were speculating that it was flat on the last round, which would have been about 2.
Edith: Flat or down.
Paul: And then it wasn't.
Edith: Bidding war.
Paul: I mean, once you look at it, it's very obvious in retrospect that the number makes sense. But at the time, I was like, "This could go either way." I didn't know if it was a fire sale, if it was fine, if it was amazing. It could have been any of the three.
Edith: This is also public knowledge. Atlassian about 4 weeks ago announced that they were taking out a billion dollar line of credit.
Paul: Oh. Interesting.
Edith: So there's a lot of chatter about--
Paul: What are they doing it for?
Edith: Well, because they took out a billion dollar line of credit just basically as a war chest.
Paul: I wonder if they were already in talks.
Edith: That's the reasoning.
Paul: You think Atlassian kicked it all off and then GitHub got some more bidders?
Edith: I don't know if they kicked it all off, but the chatter that I heard was, "What do they need this billion dollars for?"
Paul: What are they going to use it for now? They could buy LaunchDarkly.
Edith: And CircleCI too.
Paul: Yeah. A billion buys a lot of things.
Edith: So what do you think it means to the employees at GitHub?
Paul: I saw Zach Holman tweeting today who was one of the early GitHubs. And they didn't know.
Edith: But he left a while ago.
Paul: He left a while ago. Presumably he bought his shares.
Edith: I meant for the current.
Paul: Oh, the current employees? I mean, two-year run out. Microsoft salaries, the salary doubles and a nice check to kick it off. And if you're current and in a valued role you're suddenly quite wealthy in San Francisco. Not very wealthy, because it's San Francisco. But you're doing OK.
Edith: So another un-attributable rumor I heard was that the person who led the acquisition, they got their bonuses based on retention of the company they acquired.
Paul: That's a good way of doing it.
Edith: So they have every incentive to make that team extremely happy.
Paul: Once the founders make their billion, a bit more money to the founders doesn't make more sense. What's the difference between 500 million and a billion? No I mean, it's nothing. It's irrelevant. Whereas, you put that into the employees and, how many employees does GitHub have? I'm guessing 1,000.
Edith: Something like that.
Paul: A billion does a lot around there, and those people have been there longer and shorter and so on. But the 7.5 billion typically includes the retention. I'm sure that they got a healthy chunk of it.
Edith: Yeah I mean my mind boggles. Tripit, my former company. I wasn't a founder but I was early. We got acquired for 120 million which was considered big.
Paul: Holy shit, Edith. I didn't know you were rich.
Edith: I was employee 23. I made enough that I could start LaunchDarkly, but not enough to jet off to a Caribbean island in a private plane.
Paul: I started Dark on my credit card. I guess we're going to talk about that next episode. But not much money after I left Circle. Certainly haven't sold anything. So starting it off was done on the credit card. Fortunately, once you've had one company go well it's easy to raise. But I was near my limit.
Edith: A final story to end the episode. I wasn't a founder at Tripit, I was an employee. They were selling to consumers, and I was brought in as a product person selling to businesses. So I met a guy at an event and he said, "Do you feel bad that you didn't make more money off of it?"
I was like, "No. The founders took all the early risk." I was an employee. But it gave me so much to be part of that transaction because I had the money I could live off of, and then also the early Tripit founders and people who made money could put into LaunchDarkly. So our first angels were our VP Marketing and our VP Products. Or, VP BizDev who all had the money to put.
Paul: What's interesting, a similar thing happened at GitHub. Tom Preston Warner had made money at Powerset which had been acquired.
Paul: It was a big acquisition--
Edith: Acquired by Microsoft.
Paul: Oh! Powerset is now Bing.
Edith: Yeah. Everything comes in a circle.
Paul: It really does. But he had the money to start GitHub and for GitHub to be to be bootstrapped and so on as a result of that.
Edith: People say that's one of the issues with the Seattle ecosystem is that there's not enough acquisitions.
Paul: I think that's true everywhere that's not Silicon Valley.
Edith: Any final thoughts on the Microsoft acquisition?
Paul: I think in general it's going to be good for GitHub. It's going to be good for open source, it's going to be good for the space. It's definitely going to be good for the employees.
Edith: I think it's good for Microsoft.
Paul: Yeah. Oh, fabulously so.