DigitalOcean – The simplest cloud platform for developers and teams Whether you’re running one virtual machine or ten thousand, makes managing your infrastructure too easy. New accounts get $100 in credit to use in your first 60 days.
GitPrime – Ship faster because you know more. Not because you’re rushing. GitPrime helps software teams accelerate their velocity and release products faster by turning historical git data into easy to understand insights and reports.
Fastly – Our bandwidth partner. Fastly powers fast, secure, and scalable digital experiences. Move beyond your content delivery network to their powerful edge cloud platform. Learn more at fastly.com.
Click here to listen along while you enjoy the transcript. 🎧
Let’s go back to as far back as we need to around npm, and when you thought it should be a business. We understand mostly the story of npm registry, the CLI, things like that, but when did this thing become a possible business for you?
That’s an interesting question, actually. It really was a process of elimination, given the mission that I had, and the goal that I had, of keeping the registry up. The entire npm experience, the npm community - it really depends on this registry service being there to provide all the packages, and all the metadata and everything that people are publishing and installing. And we got to a point near the end of 2013 where the way that that was running was clearly not scaling to the size of the community.
It started out as just a CouchDB instance on my own personal VPS, and then this company offered to provide/donated infrastructure for it to work, and then that grew into the point where they could no longer support it, and it was starting to fall over. We had some pretty massive periods of downtime near the end of 2013, and my thought process there was, “Okay, what are the options? What do people do to fund things?”
I could fund it myself, that was one possibility, but I’m not unfortunately independently wealthy…
…and couldn’t afford to keep paying server bills. I could run it as a project inside of a bigger company. There was some talk of doing that inside of Joyent, the company I was working for at the time, running the Node project… But I could have also gone somewhere else, I probably could have talked Microsoft or Google into hiring me and making npm a Microsoft or a Google thing. There were some pretty serious downsides to that option. It’s a very precarious position to be in if you were not the moneymaker within a big company; you’re always kind of in danger of being decommissioned, or deprioritized…
[03:46] …squashed, yeah. You have to fight for resources there. Another option was starting a foundation. That was something I looked at very carefully, and really looked at very closely. The downside of that - foundations are well-suited to a certain shape of problem. They work very well for things like the Linux Kernel, the Node Project, where you have essentially no infrastructural overhead. There is some, I don’t wanna say no, but not very much infrastructural overhead; they don’t have a ton of servers. The Linux Kernel isn’t gonna stop working if they don’t have servers to service all of the Linux Kernels out there in the world. This is not the shape of that piece of technology
What they do need is they need people contributing to it, and foundations are a great way to marshal the human resource of 30 different companies all have a vested interest in this project succeeding, they all wanna put engineers on it, and they need a way to work together. So the foundation is really structured to enable that kind of solution.
The finances of how a foundation works are you go out to each of these big companies and you say to them, “Pony up a quarter million dollars this year and you get to be a platinum member, which means you get a voice on the governing board, you get your logo in very large size on the website, and you get to nominate two people…” etc. There’s different ways of structuring it.
Then every year you go back to them and say “Pay us again. You made this investment, you got this return, let’s do it again. Let’s renew.” And if they keep being happy and they keep being involved with the project, or invested in the project’s success, they’ll keep ponying up. The problem with npm is that we really didn’t need that many developer resources. What we needed was we needed to pay the server bills, and the community was growing at an exponential rate. If you look at the graph of downloads and the graph of the number of users, the number of packages, it follows an exponential curve so well you might think that we are faking it. The problem with an exponential curve is it gets bigger and bigger. That’s kind of the only thing about an exponential curve.
The trend is going up.
Right, and the bigger it gets, the faster it gets bigger.
Yeah. Compound interest kind of thing.
Exactly, exactly. Because the more people that are using npm, the more value there is in npm, the more people will decide to start using npm. And it would be a really tough sell to go to a company and say “We need a quarter million dollars for all our platinum members” and then the next year say “Good news, now we need a million dollars.” And the next year after that, “Now we need five.” It just does not scale.
There was, however, this fourth option, which was to start a company. A venture-backed company is actually uniquely-suited to handle things that have exponential growth curves. That’s exactly the kind of thing that venture capitalists get very excited about, because what that means is you have something that’s providing value based on how big it is, and the bigger it gets, the faster it gets bigger.
They’re kind of like always sniffing out for things that have an organic, exponential growth curve of anything… Even though in this case, when we’re talking about the growth of our downloads, I’m literally showing them a growth – it’s like, “Hey, our costs are spiraling out of control.” Like, “Oh, good news…” Because what it is - it’s an indicator of interest and value.
So we raised some money, I talked to a bunch of VCs in December 2013, started the company formally at the very beginning of 2014, and we were off to the races, with True Ventures putting in the first seed money.
That doesn’t mean the problem was solved, obviously. The other thing you need to have if you go down the venture route is a reasonable assumption that you can turn that exponential growth curve of cost into some kind of revenue-generating engine… And I didn’t at the time know exactly what that would look like, but what I did know was there’s a lot of big companies that are using npm; they have different needs than what the open source community has, and so there’s a good chance that we can do a thing that will help them have their needs met, and they’ll pay for that. And as the growth of users and packages and just the relevance of npm continues, then that can be a way that we can sort of tie a revenue to that growth curve and fund npm sustainably long-term.
[08:12] Had you done anything like this before yourself? Some founders come in green, they’ve never founded anything before; obviously, they’ve run some stuff, they’ve been an adult, they’ve been responsible, but building a company, building revenue, maybe even seeking a seed round or venture capital is just things people don’t often do, unless they’re in a position like you - one day you’re like “I need to somehow keep doing this, and I want to”, so as you’d mentioned, through a process of elimination you made your choice.
Right. No, I’d never founded a company before. I’d worked at a handful of startups, I’d worked at Yahoo and some other companies, but no – the only actual business-business experience I had was as a consultant, just kind of doing my own side work. So I’d done some stuff, it’s not like I was completely green on that side of things, and I had managed people before, but no, this was a brand new adventure for me; I was a first-time CEO, and again, doing it because I sort of had no other choice.
I’m kind of thinking, what was the December 2013 visits with the venture capitalists like? What did you do to prepare? Were you crazy nervous? Paint a picture there for us.
So one of my co-founders had some experience raising money - quite a bit of experience raising money at Joyent, actually - and he knew a bunch of investors. Also, the CTO of Joyent – or I don’t know if he had left Joyent already… The former CTO of Joyent, Jason Hoffman, was also giving us some advice and helping to make introductions and provide a good reference for us… But it’s really fascinating the way that the – you know, I look back over our pitch decks from the seed round, and some of them are like “Well, yeah, that is still a good idea. We’re still planning on doing that. We haven’t gotten around to that one yet…”, and other ones are just so laughably bad, I’m like “Why did anybody think that was a good idea?”
But the thing about raising seed money is you’re really trying to convince investors of a couple of things - you’re not trying to convince them that you are a profitable company; that’s more the growth phase, it’s a much later phase kind of approach. What you’re trying to convince them of is that this is a real problem, this is a legitimate approach to that problem, and “I am a person who can figure out what the solution to that problem is gonna look like.” You’re almost trying to sell the positioning of the founders, rather than the particular solution… Because any seed venture firm is gonna know that whatever thing you’re pitching right now, you’re gonna change it 4-5 times before you achieve product-market fit.
So from their point of view it’s like, “Do I think this is a legit thing?” and “Is this a space we wanna be in?”
What that also means is – and this is a controversial opinion, a lot of people probably disagree, but I think what’s in your pitch deck almost just doesn’t matter. The main thing you’re pitching is just “Here’s a big thing. I know about it. You should give me a lot of money.” That’s basically it.
You need to have the pitch deck, because it’s part of the whole ceremony of it, but the products that are included in there, or the approaches that are included in there are…
[12:07] Just to give confidence that you’ve actually thought to the problem and you have some indications of how you might go about generating revenue, basically.
Right, right. I don’t know… Any VC who really nitpicks your model that’s in your pitch deck for a seed-funded company - they’re just being an a**. They’ve already decided no, they’re just kind of trying to justify that.
So there were a handful of people we talked to who just completely left us out of the room, or were just jerks…
Some people in the VC community are not nice. [laughs] Most of them are very friendly, though. Most of the people we talked to were at least extremely professional. The biggest thing I got – there’s some folks who just don’t get how to make open source a business, and they think “How do you build a business around open source” is a single problem. And in reality, that’s like saying “How do you build a business around software?” Well, there’s 8,000 different ways, and any of them can fail or succeed in a million possible outcomes.
There is really no one way to make a business in open source, so occasionally there’s be investors who we talked to because they’d invested in a handful of other “open source companies”, but when we talked to them, they’d be like, “Well, you don’t have the Node contributors on your team, so how can this ever succeed?” It’s like, “Well, I have the entire npm contributor base on my team, because that’s me, pretty much… And this isn’t just for Node, and why would I need that?”
But people get assumptions in their head. And probably that heuristic served that investor pretty well. If you try to and create a startup around Node and you didn’t have the backing of the Node core dev team, you’d have a really hard time. You’re coming in sort of as an outsider to this open source community. In the case of npm that’s just not the shape of the situation.
Were you the “official” package manager of Node at the time? I know that was early days for npm, but were you the official…? Because it says that on your About page, and I’m just curious if that’s like a blessing thing, or if it’s just a “you’ve earned it” thing.
We had Bower, and what were some other registries at the time? Bower was after npm maybe… Definitely after, but in terms of what it was trying to to. There were a couple other competing registries; it was sort of like you put your stuff in several places, not just npm… But npm ultimately won.
[16:10] Of all the major web front-end frameworks none of them used npm in any kind of significant way other than Browserify. Webpack I believe still didn’t exist yet. I could be getting my Diesel mixed up, but it certainly wasn’t the thing it is today. And yeah, that was a big open question at the founding of this company - how do we get front-end developers to put their code in npm, and how do we get them to consume their code from npm?
As it happened, we just always had bigger problems to solve, and a couple years in we kind of looked back and we were like, “Oh, I guess we did it.” The answer was “Continue existing and wait.”
Right. “Be reliable” I think would probably be another one in there, because there’s been some - and I’m sure this is probably a thing that keeps you up at night - ups and downs that a registry can have.
The speed… The global speed even, not so much just here in the United States where we’re at, but Europe and China and other places.
You know, my number one startup maxim is just “Stay alive long enough to get lucky.” I think there are very few cases where success is not 99% luck, but you do have to be there to get lucky, which means whatever the fires are, you have to put them out fast enough that when you get some lucky break, you’re in a position to grab it.
So it’s not that there’s no skill in getting lucky, but it doesn’t mean you’re a genius; it just means you managed to be present and alive and still a relevant concern. If you go out of business, then you can’t really get lucky anymore.
[19:33] Did the name ever have an issue? I know that there’s always been a constant debate of what npm actually means… Because I do believe in episode #101, when we first talked to you - this was August 2013, on the Changelog, “npm origins and NodeJS” was the title of the show… I’m pretty sure we called it “Node Package Manager.” I’m pretty sure it was called Node Package Manager back in the day, but since you’ve kind of zoomed out a bit at a macro level and de-named it, and now on the top bar of your site you even – I’ll just hop over to the next tab and see what it says right now…
We sort of joke about it, yeah.
Yeah, I mean… Because at first it was Node Package Manager, and I think now it says naughty I can’t pronounce the middle word something made. I don’t know that’s what it says on your site right now.
Well, that’s randomized; it’s just for you.
Well, let’s turn to that. Let’s break it down first, the components of npm. You do it well on your About page, which I really appreciated, by the way - just sort of breaking down what npm Inc. is, versus npm the registry, versus npm the CLI client that kind of pulls everything together, stuff like that. I broke that down I guess for the audience, but what are other things that you’ve done that are paid for, that generate the revenue? When did you first launch a product, when did you first start earning revenue and what was that like for you guys?
[23:49] This is an interesting story… I’d had people from various companies come to me and say “npm is great. We love npm, we use npm all the time. Hey, here are these three problems we have.” And these weren’t small companies, they were big, well-known companies with lots of money. That’s one of the reasons why I started the company, because I was like, “Well, this is a way that I could fund the registry and keep it running.”
One of the things that we heard in a lot of our very, very early beginning customer chats was people seemed to split about 50/50 between absolutely not being able to use anything that is not on-prem, and absolutely not being able to use anything that’s not a SaaS. So we thought, “Okay, we’re gonna have to have some kind of a two-pronged attack here”, where we have an enterprise product which is on-prem, and like more of an SMB type of thing, which is a SaaS product. And we looked at GitHub - that’s kind of the obviously corollary, right? GitHub had GitHub Enterprise, which is run on-prem, and GitHub Orgs, which is their paid SaaS for teams and smaller groups.
So the first thing we built, just because we found our first customer for it, was this on-prem npm enterprise product. I would not recommend that any company today create both an on-prem product and a SaaS product. I don’t care how well-intentioned you are, I don’t care how much you say it’s not gonna happen to you, it will.
A lot of people told me “It’s important that the codebases don’t diverge”, and we were like, “Okay, let’s keep the codebases in sync.” What they didn’t say is “No, no, no. They will. They will diverge. You can’t stop it. It’s just going to happen, there’s no way out of this…”, and the reason why is that it takes two completely different kinds of company to make an on-prem product work, versus making a SaaS product work, and you really just have to pick one and go with it.
An on-prem product - the big challenges are onboarding, customer success, managing upgrades… The whole account management process, and just feature delivery. It’s really challenging to get any kind of upgrades or features out to a customer; you have to convince them to do an upgrade. You have to make them successful, you have to build in the kinds of corporate competencies to handle that kind of a customer need. So it’s not just about having a good support team, it’s about having really proactive, technical account managers who can reach out and be like, “Hey, how’s it going? How many people do you have on this thing now? Okay, we’re gonna need to upgrade your server. Let’s take a look at your CPU logs and make sure you’re not getting yourself into trouble.”
What happens with npm is that the more you use it, the more you wanna use it within a company. So people would spin un a proof of concept with 20-30 developers at their company, and then expand that out to a couple hundred or thousands, and then it would fall over; and it’s like, “Yeah, I’ve seen this happen.” This is 2013, it’s like a microcosm for what the open source community went through between 2009 and 2013.
At every company we went to, this was the same story every time. So you can do that. You can manage that well, you can do that with a lot of really good, proactive, professional services, and customer success, and so on… But that’s like a certain kind of team that you need to build.
On the other side you have SaaS, where you really benefit from extremely rapid delivery, continuous deployment, metrics to be able to determine when customers are running into problems before they do, so that you can get fixes out. It’s extremely easy to push new features out; you kind of just push it out, and then there it is… And the big problem is doing the operations. You’re not in a position where you’re facilitating an ops team at another company; you’re in a position where you are that ops team, so you need to build that technical competency.
Those two companies are diametrically opposed in terms of how they’re actually built. You can do both once you’re a 200, 300-person company, because you have enough hands to actually manage both approaches, but doing that as a seed stage company is just impossible. It was a fundamentally bad idea that took us a long time to unwind from.
I’m curious with that then, so the state of those two - so you do have enterprise, and you do have your SaaS, so what did you do?
[28:13] That’s a good question… We haven’t’ made a ton of noise about his; it’s not in GA yet, but what we did was we went back to the drawing board… I had some revenue targets that our investors wanted to see; they’re saying like “Your revenue should be growing about this much, over this much time…”, so I did the math and I was like, okay, we can do that if we land – given the typical annual contract value of our enterprise clients, if we land about two new enterprise clients per week over the next year, we’re gonna hit that target.
I could hire a sales team that could do that, we could invest in the marketing to do that… It’s not easy, but it’s not particularly challenging either. The what-to-do is very straightforward.
Right, you’ve gotta do the work.
It’s the main thing!
Right…? You have to keep the main thing the main thing.
The moral of that story was “Let’s start looking at how we’re gonna go to the enterprise in a different way.” So we reimagined our enterprise offering as a SaaS, and basically what it is is the public registry and the public website, with all of the features that those have, but sort of in miniature, and rolled up into a Kubernetes cluster that we run for the customer. This really leverages a lot of our strengths; we’re very good at scaling an npm registry. We’re probably the only team in the world capable of scaling an npm registry… At least the only one that’s ever done it at this magnitude. And also it’s just a lot easier for the whole sales cycle.
We were also wading into these companies with this minefield in front of us, where on the one side we have the developers, who are bringing us into the company, saying “We need npm, we need this tool. It’s how we do our work. Collaboration is getting really hard across the company lines; we need something like npm.” And the infrastructure teams at that same company, they saw us as the enemy more often than not, because what we’re doing is we’re making work for them. They’re not getting any benefit, really, from having npm. What they’re getting is yet another thing to kind of care for and feed. It’s often coming out of their budget; they have some limited tools budget and they have to deliver what the development teams need… So their pushback was always like “Well, can’t you use this other thing?” The dev teams would say, “Yeah, we could, but it’s not as good”, and they’d say “Okay, great. You’re gonna get the thing that’s not as good then.” [laughs] Because the last thing that they wanted was some new thing to be aware of and manage and worry about scaling, and in fairness, we weren’t able to provide them with enough support to be successful at that… You know, back to the whole SaaS versus on-prem thing.
So reimagining this as a SaaS and redesigning it as a SaaS actually opens that door up a lot. Now the infrastructure and operations in the IT side of the companies is saying, “Hey, you just took a big headache away from me.” The dev teams can get started a lot faster, and in a lot of cases there’s nothing to install, they don’t have to requisition any hardware, so they just swipe a credit card and off to the races.
So it’s in beta now… We have gotten some pushback from some companies about it not being on-prem, but the interesting thing here is kind of the subtle repositioning of like “This doesn’t have to be where you put your private code, if that’s a problem for you to have that in a SaaS repository.” You could just look at it as a way to kind of enforce policy compliance within your company, because you are using a SaaS, right? You’re using the public registry. I guarantee it.
We can look at our logs and find 500 of the Fortune 500 companies. People think you’re lying when you say “100% of the Fortune 500.” They’re like, “There’s no way that’s true.” No, it is actually true. It’s not 99%. There aren’t five that aren’t using it; it is 100%. Given that, then the thing is “How do we help the policy and compliance teams? How do we help the IT and the security teams?” and not just sort of give developers what they need… Because that’s our bread and butter; that’s who we identify with and who we’ve always served. In the open source world, that’s really the only person to serve. But looking at things through the lens of security and policy compliance, you start to see this as a bigger equation, and there’s a lot more opportunity to offer value by thinking a little bit more holistically about everybody within a large organization.
That’s pretty smart. If I understand you, the current state of enterprise is moving towards a SaaS-based version of it; it’s in beta. Prior versions were a little separate project/codebase that was installed on-location, and then you had sales teams and technical teams kind of like paying attention to them, and monitoring it, or whatever, to keep them getting the value that npm on-prem would bring.
It’s also interesting to look at that from the world of cloud, too. We’re trusting someone else to host and run and operate your codebase, or at least infrastructure that runs your code has become –
It has become commonplace.
Yeah, exactly… Whereas before you were sort of forced to go on-prem, because the trust wasn’t there. But now you’ve got it, so you’re able to do enterprise in a much better way, that doesn’t divide your company.
Right, right. There is an interesting conversation we had around what that trust means, and if you look at the data and the research on this, in many cases and for many companies, unless you’re Microsoft or Google, you’re probably better off trusting Microsoft or Google to run your infrastructure, rather than try and do it yourself. Even if you have a very capable team, they’re just not operating at that scale… So you get a really big business advantage by going with the SaaS, in a lot of cases.
Now, that being said, a lot of times when people say “on-prem”, they don’t mean literally on-prem. “On-prem has come to me in our AWS account, rather than yours…” There are some hold-outs, and I think as our enterprise go-to-market kind of progresses, at some point we’re gonna find someone who says “No, no, really… I want this thing. I know it’s this cute Kubernetes cluster thing, but I’m gonna pay you many millions of dollars to have it on my physical location.” Then at that point we’ll say, “Alright, let’s build the team that can support that.” But we’re now in a position where we can go to most enterprises and sell them a solution that makes their lives profoundly better for everybody across the company, and we don’t have to get into that fight between dev and infrastructure. We don’t need to ask our champions within a company to go convince someone who sees no personal benefit to using our product.
[36:19] So our existing enterprise customers – I wanna mention this just in case any of them are listening to this… [laughs] We’re going to keep supporting our npm enterprise on-prem product for quite a while. It’d be a little messed up to say “Okay, we’re dropping support for our on-prem product, but our new SaaS product is not even out of beta yet.” So as we do get closer to rolling this thing out and calling it GA, we’re gonna have a comprehensive plan to make sure we take care of those people.
Well said, yeah. In case they’re listening, you’ve gotta make sure that they don’t get “Hey, I should be emailing Isaac after I listen to this show, because that was not expected.”
“Did I just hear you decommissioning my thing?”
No, no, no… It’s cool, it’s cool. We’re good.
Right. Well, let’s turn back to the – I don’t wanna get too far back, but just enough to kind of understand this… So if I understand you correctly based on some of my research, you took an initial seed funding in 2014, in the February range, at least based on announcements… Literally one year later you took another series A for eight million, so you’re looking at maybe 10.6 total of funding… And you haven’t been funded since then. Does that mean you don’t need more money, or you’ve been just so wildly profitable? What’s the scenario there?
There is a third possibility you didn’t mention, which is that we got money and didn’t say anything about it.
Oh, is that right? Okay…
How much money did you get and why didn’t you say anything about it?
Well, since we didn’t say anything about it, I probably shouldn’t say anything about it, but… From the point of view of looking at this as a founder, you should use funding announcements very carefully. It’s certainly not required to announce every dollar of funding that you get, and it often doesn’t really serve your company’s needs.
I think that a seed funding - yes, announce it. Announce it as soon as possible.
Exactly. A funding announcement is always marketing. If you have a really big round, then you probably wanna announce that; again, because it’s marketing. It can rattle the cages of all your competitors, it gets all your buyers thinking, it gets you in front of a bigger class of buyer, who’s now thinking, “Oh, they’re a significant company, because they’ve raised 50 million dollars”, or whatever. But if you do follow-on rounds, if you do like a A1, you probably don’t need to talk about that… Especially if you’re doing inside rounds, or anything like that; it doesn’t move the needle in terms of taking your company seriously. It’s not the best way to brag about your progress, and it doesn’t really get you anything to make that announcement.
It makes sense.
So without saying too much more, you know, if you’re sitting on the feed of SEC filings or whatever, you would see that we have raised some more money, but nothing that we’ve made any kind of big news about. Honestly, I think hiring a new CEO is a much bigger deal than any round we’ve raised.
That’s where I was trying to go to, because I wanted to kind of close the gap there… One, to give you a chance to say “Yes, we’re so wildly profitable”, if that was the case…
Because that was at least what I was probably trying to gather from the total – what it seems, publically… Total raises, and kind of painting a picture or a position for npm, prior to moving on to the next – I would say a much bigger topic is hiring a CEO, and what that means for you… Because a lot of the story we have covered, you’ve been not only leading product, but also leading as a CEO, which straddling those two lines is super-difficult.
[40:03] Something you had said actually in your announcement post was basically you said being – I’m gonna go back to where… You said “One piece of advice I got as a founding CEO of a startup was to start planning your transition and less to be planned for you, so… It seems like this is a big deal for you, obviously. You’ve personally penned a post deeply about it, which I’ve just quoted from… So help us understand why this was such an important stage for npm.
Sure, sure. Yeah, people really responded to that post. It’s one of those things where I just kind of was like, “Oh, I should say something”, and I just kind of wrote a rough draft, and… I think one of our marketing people was like “Oh, this thing kind of sounds weird. Maybe restructure that sentence”, and other than that, it went out as is.
Yeah, why did I hire a CEO…? [laughs] No, I really like him. I genuinely like the person I hired. I think we lucked out. Bryan has been great to work with, and for, and… It’s kind of a weird situation, as a founder, to hire a CEO. On the one hand, you’re hiring your boss, but on the other hand, I’m still on the board. I have all kinds of shares, and a ton of credibility, so it’s almost like he’s working for me… But it has to be a very good partnership, and it has to be something where you have a lot of trust in each other.
He has to have trust in the company and in what I’ve built and what I’ve done so far and what my vision is, in order to get on board and wanna take his own time and energy and spend it on this. And I have to really trust that he’s not gonna screw it up. It’s my baby, and you’re hiring somebody else to be your baby’s new parent.
Right… Very tough to hire for. That’s why, the larger part of the conversation – actually, this wasn’t the most I wanted to talk to you about, but this is hard, right? And someone listening to this will hear your passion; you’ve been in the known world since the earliest days… Core contributor, Joyent… Your full story can be heard on other shows; we’ll point to them in the show notes, so go back and listen to other shows for that to kind of get some of this back-story if you want to, but… You know, hiring a CEO as a co-founder, as someone who’s been in charge, has got to be an extreme exercise of not only trust, but maybe what I’d say is responsibility… I’m not hitting the right word, but a very adult thing. It shows that you don’t take yourself so seriously that you have to have that title, have to have that role.
But you do take yourself seriously, because you do run things, of course, but you’re not putting Isaac on such a huge pedestal that no one else could do the job.
Yeah, yeah. I mean, I had a few sleepless nights over it, to be honest, over just the idea or the concept… And it wasn’t even ego; it was just like, “How can I do this, and how can I do this in a way that’s not gonna be absolutely terrible and take this risk?” Ultimately, where I kind of got to was like, “Okay, yeah, it’s a risk. If I hire somebody else, they might not be good, the team might reject them, we might have this massive upheaval…” We probably will have a massive upheaval no matter what, just because it’s a new person who’s gonna come in and be like, “Hey, all these processes you set up - this why you’re failing. We’re gonna change them”, and change is disruptive and traumatic, right? There’s no way around that.
But it’s almost like the company equivalent of a heart transplant, or something. It’s like, “Yeah, it might go good, it might be life-saving, but it is not a small thing. It’s painful and big.” And there was some pressure from our investors… I think some of our investors were like, “You know, you seem to be doing an okay job, but here’s what you need… Maybe you need to hire a COO, or hire somebody to run marketing, or somebody to run sales…”
The hockey one was good…
[laughs] Okay, good. How do we make sure that we’re well-suited to handle the changes that we’re causing?
You need to be in that position. That’s the job that no one else can do. Sure, there’s some others who could probably do it, but you probably have a better capability of doing it well.
Sure, sure. I have a track record of doing okay at that. And that’s not like – I’m not trying to be modest or bragging about any of this; it’s literally just like stepping outside of my own skin for a second and looking at it as a shareholder, like “What does this company need? What is the mission of this company and how do we serve it?”
And would I hire this Isaac guy as a CEO? The answer is “Hell no.” He’s a smart guy, but he doesn’t want the job, so… That’s kind of a deal-breaker.
That is so funny, to say it from that perspective… “Hell no.”
Right. So we did this first CEO search beginning of 2017, first half of 2017… And everybody kind of fell into a couple of different categories. There were some people who we really liked, who were very well-qualified, and said “You know what, you’re not ready.”
As a company.
Yeah, as a company, as a product… One person in particular, she kind of gave us this “It’s not you, it’s me” kind of comment, and I drilled down into it and she was like “You know, you need somebody with more developer empathy, who can figure out what your product vision needs to be, and I think that needs to be you. Until you do that, you can’t hire a CEO.” And I was kind of feeling like, “Well, as long as I’m CEO, I can’t do that.” [laughs] So it really was a frustrating time, in a lot of ways.
There were some other folks who really wanted the job, and as we went through the interview process, and especially the background process, I was like, “Okay…” I could see that any company this person works at is gonna be successful, and they’re really smart, but I’m gonna be budding heads with them every single day. Do I want that to be my job? The answer is no, I don’t really want that to be my job.
What I didn’t wanna do is hire somebody who’s gonna go very effectively sell this on-prem product that was sort of becoming toxic within the company.
It’s the box, from Silicon Valley.
[47:55] [laughs] Exactly, exactly. I didn’t wanna reimagine the entire company around this approach, this product approach that I fundamentally no longer believed in… I was like, actually what we need to do is pivot and change our direction a little bit, and focus on this other direction. And then there were some folks who I think really did get it, but were kind of on the fence… And yeah, again, I think a lot of it is like – you know, they were looking at this and saying, “Well, I can spend my time figuring out what your product needs to be, or I can spend my time running your company, but I can’t do both, so… I’m gonna go somewhere else.”
So we did a six-week – we actually commissioned one of the people who passed on the CEO search to do a six-week market research and marketing customer research project with us, and that’s kind of where we landed on like “Look, we need to get into security, we need to be a SaaS, we need to be the thing that serves the interest of the IT and security teams at enterprises, rather than just the thing for developers.” Because are uniquely situated to be able to offer a security tool that doesn’t cause more friction for developers. And that I think is a pretty unique opportunity, a pretty unique thing in the dev tools space.
Well, given your distribution too as a registry, I think the markets also said “Hey, you need to have some security as well.” While it has not been your fault, you’ve been in the vector of breaches.
Of course, yeah. We’re in the toolchain, we’re in the supply chain of your software.
Right. You need to be a check and balance.
You’d said part of – how do you pronounce his last name? Is it Bryan Bogensberger?
[laughs] Right, right…
npm is actually really interesting, because there are opportunities to provide very, very compelling enterprise tools that don’t ruin our community, that don’t actually harm the open source users. There’s not a ton of open source companies like that. There’s usually some kind of contention, especially if you’re talking about an open source project that’s delivered as a standalone executable. You kind of have one classic example of this, the way that MySQL was dual-licensed for a long time. I forget now… There was like MySQL and then MySQL AB, or something, or MySQL BC… Anyway, there was the commercial version and the open source version. The open source version was GPL; if you want that non-GPL version, you have to pay money.
That’s a lot trickier to do in a way that’s not very fragile. For example, what ended up happening with MySQL is that it was bought by Oracle, who just stopped updating the free version and said, “No, you’ve gotta pay.” And a lot of the open source community had to just go elsewhere, and you see other things… MariaDB is the main one that all the MySQL folks jumped to.
The other thing is you can build an open source version which is free, and then add features to a commercial version. I remember there was some talk about doing this in Node back at Joyent that I resisted very forcefully at the time, which is that we would have the Node open source platform, and then there’d be another Node commercial platform that would have all sorts of other APIs or other things built into it. And I think when you’re in a position of running an open source project, doing that is very, very difficult. It gets you into a very difficult ground, because you have to walk a very fine line between really having this value and then holding it back from your open source community, which to me just feels like – not like there’s anything morally wrong with it, but it just feels gross to me.
I feel like I have this thing, I’m giving it away, but I’m not giving you the whole thing, and there’s no real reason why one part should be included or another shouldn’t. Now, at Bryan’s former company, Inktank, that’s exactly what they did, and I think they walked that line relatively well, to their credit… Where Ceph, the storage platform, was completely open source; it was liberally licensed, they had tons of people using it, and then all of the management infrastructure, the line that they drew in it - I don’t wanna speak for his former company, that I know very little about actually, because I never used Ceph, I was sort of busy with other things, at the other end of the stack, throughout that whole time… But yeah, there was a pretty clear line in the sand between what is commercial, what is paid for and what is free for the open source community. But there’s always gonna be those edge cases.
With npm, we don’t even have to hold anything back. The open source tool can have 100% of the functionality because that’s the client to the registry.
It’s in your best interest that it be that way.
Absolutely, absolutely. And it’s actually competitive, because now people who are providing alternative registries, who are competing with us - they have to either seem a little foolish, because they don’t support the whole feature set, or they have to be spending time and energy scrambling to keep up with us, and we’re sort of always fundamentally in the lead, and really taking advantage of our privileged position there.
[56:10] But the other thing about it - there are features that open source users fundamentally don’t need out of their registry service. They don’t need private packages, for instance. They don’t need the single sign-on, and management and policy compliance stuff that you’re gonna see inside of really large enterprises.
There are certainly some arbitrary lines to be drawn. I think GitHub just recently started giving away private repositories for unpaid users, which I think is interesting. They limit the number of repositories you can have. All of their free users get now three private repos, and I can see us doing something similar to that, and continuing to evolve where those lines are drawn, but fundamentally, the needs – there are needs that a large enterprise of 1,500 developers… They absolutely need a way to say “Here are 25 different projects, and I need to see what’s going on in all of them.” Nobody needs to know all of the things that are going on in all of the open source projects. That’s just – you couldn’t, a. And b, it’s not really reasonable to want that.
We can build a lot of really compelling features that are structured in a such a way that we’re… As one example - we haven’t built this is; this is vaporware, a handwave on the roadmap, blah-blah-blah… But for example, whenever you run npm install now, we run an audit of all the packages in your dependency tree. That’s something that benefits every single one of the people using npm today. We give that away for free, and why do we do that? Because one of the things that you need in an enterprise is you need to know how are all of the developers at my company – have any of them installed something that is insecure? Have any of them built an application and shipped it to production that is insecure? Can I go and verify all of the different licenses that are in use in all of our different projects? And these are the kind of things that that same auditing infrastructure enables us to do.
So we’re in a position where giving things away to the open source community actually does serve our long-term product and company needs as we go to market in the enterprise.
Something that I read on your About page and I actually wanted to confirm with you if this is still accurate - I imagine that usually most people’s About pages are the last thing they look at, but it’s one of those things where you seem like the kind of person to be very intentional… So I’ll read it real quick. It says: “npm’s mission is to take open source development to entirely new places. When everyone else is adding force, we work to reduce friction. npm is not a typical product, and we’re not a typical “Work hard, play hard” startup. We are responsible adults, with diverse backgrounds and interests, who take our careers and our lives very seriously. We believe that the best way to iterate towards success is by taking care of ourselves, our families, our users and one another. We aim for a sustainable approach to work and life, because that is the best way to maximize long-term speed, while retaining clarity of vision. Compassion is our strategy.”
It also said somewhere else too that you just really care about the community, and I think it’s important to look at npm in this way too, that you run the registry, there’s some vested value there for you as a company, but also the fact that you run it freely, scalably, fast, for the globe, the entire world, to me says a ton about you as a company. And to find ways to do what you’ve done just there - bringing security audits to every user, not just those who pay for it - to me says how much you care about the community. But… You need to make money as a company.
Well, I couldn’t have asked for a tougher question there, buddy. [laughs]
[01:00:13.05] So that thing in our About page - that’s an interesting story, actually. I wrote that as part of our first hiring blurb… When I wrote that, I believe Laurie and I had just gotten done dealing with some huge operational issue, back when he and I were the two technical people at the company. And we were sitting around the table at our VC’s office, at True Ventures’ office in San Francisco; we didn’t have our own office space yet, and it was like 8:30 at night, and it was like…
Yeah. “We’ve gotta hire some people who are not gonna put up with this stuff… Because this is not gonna keep working.” So it was a lie that came true… I guess aspirational is the best way to put it.
Yeah. But we’ve assembled a world-class team here, and I think actually in particular if you look at age distribution at most startups, the average age at most startups is like 30 or younger, and I think at npm we have actually a pretty representative, or even skewing kind of older, in terms of the people at this company. There’s a ton of folks who have been involved with open source for 20-30 years.
On the one hand, we don’t ask people to work stupidly long hours. We also do kind of trust that everybody’s gonna manage themselves appropriately, and not fry themselves out, but that does happen… But yeah, we do take a very intentional approach to how we build things, and how we set priorities, and try to be very realistic about what we can deliver, and cognizant of what our needs are.
So is it still true?
I think so, yeah.
Is there anything about that you would rewrite, or change, or add to?
I don’t know, that’s a tough question. There’s probably some stuff on our About page which isn’t correct. I haven’t actually looked at it in a really long time.
I only pulled it up because I actually liked it quite a bit, and I know that the usage of an About page from company to company varies. Microsoft’s About page, versus npm’s About page probably skew differently, right?
When I read that, it reminded me that, hey, when I looked up on LinkedIn, you’ve got 60 employees or less, where you’re trying to go, at least - I’m not sure if that’s accurate or not, but… It reminded me the utility that the npm is, but also the company that you are, and the position you’ve been in, so in preparation for having this conversation asking you about “How do you hire your new boss when you started the company?” is gonna be a tough question, for one, but then what a responsibility it is to the community that you come from this position, this perspective.
I think it’s important.
[01:04:00.07] Let’s close with this one then - what’s in the very near future? What’s on the horizon for you that people may not know much about, that you can at least tease or share some general details publicly now? What’s on the horizon, since this is a transitional period for you, you’re moving into the role of Chief Product Office, spending a lot more of your time focusing on the things you love more of…? Where are you going?
We are hiring… We’re hiring pretty aggressively, I should say. We’ve already grown the company – it’s been ridiculous…
Was 60 an accurate number on LinkedIn?
Probably about right, yeah. This is another interesting thing as a founder - and not as a CEO, I should say - we’re just now at the point where this last batch of people that we’ve hired a couple months ago was the first time somebody showed up in Slack that I had never met… And it was just like “Oh, who are you? You work here? Huh, that’s novel… I was not a part of your hiring process. How did this happen?” I kind of had this weird, like, “There’s a stranger in my house” kind of feeling at first, the first couple times it happened.
Right, and have to trust them in the refrigerator.
Right. I’ve gotten used to it now. I try to set up meet and greet chats with everybody, with all of our new employees, every so often… But yeah, we are growing really quickly, we have just a team that’s exploding. We’ve built out a whole new executive leadership team, added a head of sales, head of community marketing, a COO, you name it.. We’re gonna be talking more about those individuals over the next couple months probably, making some noise about them… And our main number one focus right now is just kind of getting to this GA launch of npm Enterprise, which should be right around the corner here, a couple months.
Well, Isaac, I’m proud of you. I think what you’ve done is great. The word I was looking for earlier in our conversation, I couldn’t find it when I was saying “responsibility” and “adult” was “maturity”. I think you’ve made a mature choice in understanding your limitations, but at the same time understanding the needs of npm… And not only the company, but those who work there, as well as those build wealth and thrive and serve an awesome world and do awesome things in software because the npm registry is trustable, it’s reliable, it’s secure, all these things. I think you do an amazing job as a company and as a person…
I was excited to have this conversation with you and dive into your past and learn more about where you’ve been, and the bloody knuckles you’ve got, and the scars, and the lessons learned, and things like that, so… Thank you so much for your time, I appreciate it.
Our transcripts are open source on GitHub. Improvements are welcome. 💚