Gone are the days of quarterly or monthly software update release cycles; today’s software development organizations release updates and fixes on a much more frequent basis. Edith Harbaugh just wants to give business leaders a modicum of control over the process.
The CEO of LaunchDarkly, which was founded in 2014 to help companies manage the software release cycle, is trying to reach customers who want to move fast but understand that moving fast and breaking things won’t work for them. Companies that specialize in continuous integration and continuous delivery services have thrived over the last few years as customers look for help shipping at speed, and LaunchDarkly extends those capabilities to smaller features of existing software.
“What we like to think is that we make continuous deployment palatable and realistic for business in terms of fitting back into their own processes,” Harbaugh said in a recent interview with Protocol. The company calls it “feature management,” and customers and investors seem impressed: LaunchDarkly raised $200 million in new funding last August that valued the startup at $3 billion.
During our conversation, Harbaugh discussed LaunchDarkly’s IPO ambitions, how feature management grew out of a media campaign and how it balances operating on the cloud while facing potential competition from those same cloud vendors.
This interview has been condensed and edited.
You’ve been raising a lot of money this year. Where does that put you right now, and where does it take the company over the next few years?
I co-founded the company seven years ago; well, it's eight now, in 2014. And it was a very new market back then. Nobody was offering, really, a commercial solution for what we were doing for feature management.
What's happened over the last few years is just incredible momentum. We have over 2,000 customers worldwide; we have marquee customers, like IBM and Atlassian, who rely on us. And we have 22 of the Fortune 100.
In terms of the public markets over the next year, is that something that’s on your radar?
We brag about having 2,000-plus customers, but Atlassian has 50,000-plus. So I think there is still a lot of growth for us.
I think an IPO at some point is on our radar: It's not imminent, for lack of a better word. And also, my co-founder came out of Atlassian. And it's funny, because their market cap is like, 10x since when they went public. We look at it as just a step, not a final destination.
How hard has it been to get companies to change the way they think about deploying software? The big tech companies, the big software companies got religion on releasing smaller changes as quickly as you possibly can. I'd imagine that's a little bit of a harder story to tell outside of the core Silicon Valley companies.
We try to meet people where they are. It's terrifying if you're a traditional company, and you're like, I'm gonna go from a quarterly release to [a release] every hour. And [maybe] I don't actually want to do that with my business, I might be in a regulated industry, I might have a union where they need to get advance notice of a release.
We help them on that journey by saying, "You can release when you want, and that's what we'll help you with. You can release when you want, and then put it out to your customers at a timetable to make sense for you."
So a real win is if you're a bank, you can say, "Hey, I'm gonna push this major update," and have it out. But then maybe just turn it on to branches in New Jersey and make sure that it works before I push it out to the entire New York metro region. That’s a really classic use case there. What we like to think is that we make continuous deployment palatable and realistic for business in terms of fitting back into their own processes.
I was actually wondering if you consider yourself a continuous-delivery company, because it almost sounds a little bit different.
What we do is we empower developers to build features, and then we empower the business to release them when they want. I have been working in software for 20 years, and I get so excited about all the use cases we help with.
I used to be a product director at TripIt, and a classic problem we would have is when we wanted to do a big press launch. And then we would have to time the deployment of a feature with the press launch. What you can do with LaunchDarkly is you can push stuff out, and it's ready to go. And then you just flip a switch, and it's live, which is incredibly powerful. And it's continuous delivery in that developers have that power, but it's also at the pace that the business needs it to be at.
Is there any kind of a pushback brewing in terms of a slow software movement? There were lots of good security and feature reasons to accelerate software release schedules, but I sort of wonder, everything tends to move in cycles, right?
Something that we talk to our customers a lot about is that we work at the pace that you want to work at. For example, we have airlines that want to do a release at a scheduled time, around when it makes sense for them and to also roll out training to maybe thousands of people. What they like about LaunchDarkly is that they can then bundle together various features, and have them ready to go again, and then just flip them live, instead of that nerve-racking thing of, “We're going to push out all these changes at once without anybody ever having seen them.”
It's also a really common use case we see with educational software. It's extremely disruptive if you're a teacher to suddenly have everything changed on you when you're trying to teach a class. So they really like LaunchDarkly, because they can give it to a couple teachers at a time that they know about and get some feedback. And once it's really solid, push it out to the rest of the teachers with a timetable that they're aware of.
I really want to emphasize that it's when you want to. Not do this artificial timeframe of, you know, YOLO , and just push garbage.
One thing I always ask folks like you who are running companies that depend on cloud services is how you think about the potential competition from cloud services. A lot of them are looking at offering services like this to their own developer communities, which are obviously pretty massive.
Both Amazon and Microsoft have been very good partners. We actually have a really successful partnership with Amazon right now where people can use their AWS credits to buy LaunchDarkly. It's gone very well so far, I think of them as just somebody who's helping the ecosystem. Same with Microsoft.
Amazon actually announced a competitor , it's funny, during re:Invent. It just doesn't seem to have gotten much traction.
