Kyle Mathews is the founder and CEO of Gatsby, a new company he’s building around an open source project of the same name. Gatsby as a project describes itself as a flexible modern website framework and blazing fast static site generator for React.js. At the macro level — Kyle’s career has been focused on a better way to build and ship websites. It seems he’s done just that with Gatsby’s launch in late May 2015…since then he’s taken on a co-founder and a seed round of $3.8M to form Gatsby Inc.
Linode – Our cloud server of choice. Deploy a fast, efficient, native SSD cloud server for only $5/month. Get 4 months free using the code
changelog2018. Start your server - head to linode.com/changelog
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.
You know the feeling of getting started. It’s a lure for us all, it’s something new, it’s something that hasn’t been done before, or at least not the way you do it, so it’s exciting… But startups are a little bit different, especially ones that are started with seed funding or venture capital’s involvement, and especially startups built around open source software. So naturally, that’s where this conversation with Kyle Mathews, the CEO and one of two co-founders of Gatsby got started.
So how did you know it was time to build a company around Gatsby?
I mean, you never know something like that… I had kind of the inkling. I’ve done startups - I’ve either worked at startups or started my own stuff for pretty much my whole adult life, so I had a pretty good inkling that Gatsby could be a startup… But in the past I’d just kind of jumped into stuff when I was very underbaked, and with Gatsby I very deliberately just told myself that I wasn’t going to start a company until the evidence was very strong.
For Gatsby, in this particular case, what that evidence was was basically – first, I went full-time on Gatsby; I just worked on the open source, it was not a company yet, two years ago. So evidence one was just that I was able to find contract work around building things with Gatsby, or even people paying me directly to make improvements to Gatsby, and I was able to pretty consistently find as much work as I wanted during that time. That was a strong indication that there was a lot of commercial value going around Gatsby.
The second big indication was just the excitement and pick-up around when this was getting closer to launch and when it launched. There were a ton of people using it, a ton of people writing PRs, so it was very clear that it was getting very serious and sustained usage… Which is also an indication that a company could be built around it. And also just that there were some big companies that were reaching out, saying like “Hey, we have serious problems with how we build websites”, and Gatsby and React, and the other parts of the stack that was developing around Gatsby solved these really well. So this was another strong proof point that the Gatsby ecosystem could sustain a large company.
[00:03:47.23] I think the interesting thing that begins to become more clear, and what I like to examine in particular on this show - to you and I it may seem like a no-brainer for a company to be built around open core or open source or however you wanna frame that, so I think there may be people coming to this kind of show with potentially a different lens or different experience level of “I thought open source was free. I thought that you just give this stuff away, and somehow, magically, [unintelligible 00:04:11.27]” which you and I both know that that’s not true… It’s like, “Here’s this crazy guy getting 3.8 million dollars from investors that have done some tremendous stuff, and now he’s forming a company.” To you and I, that may seem kind of normal, and maybe even in Silicon Valley that’s normal, but I think it’s becoming more and more normal, but to many it’s just not. How do you rationalize that being normal these days?
It’s definitely not normal in the sense that it’s common.
It doesn’t happen every day.
Yeah. I think the disparity is that most people’s experience with open source is definitely non-commercial; you throw something up on GitHub, and sometimes people put up issues or even submit a PR, or you submit PRs to other things, or whatever. But there’s a big spectrum of open source projects. They go all the way from very simple, like 50 lines of code, that encodes a neat little algorithm that’s useful sometimes, up to incredibly complex projects that are like core infrastructure for thousands of other projects.
Most open source projects definitely cannot support or sustain a company around them. They’re just not valuable enough. Well, there’s two problems - they’re either not valuable enough, or there’s not an easy business model that can be built around it. If either of those two conditions don’t apply, then you can’t build a business.
You can think about it also from a perspective of – so businesses exist to a) create value; so there’s value being created some way or another through the activities of the business, and then the business sustains itself by capturing some of that value; generally speaking, customers pay for something. So for open source to work – open source businesses tend to capture less value than purely proprietary businesses, which is why open source is so great; there’s so much customer surplus being created. But for open source to work, it has to create a ton of value, because it’s probably capturing less than an equivalent proprietary business, but it also just has to be able to capture value at all, which is a sticking point for a lot of open source projects… It’s like, “How do you actually capture some of the value that’s being created?” Some types of open source projects just don’t lend themselves to that, in which case you’re reliant on volunteers, and maybe some companies that it’s important enough to them that they just sort of sponsor part of the work done in the project.
Maybe to give a little timeline here - in 2015, late May, you launched this as open source… I wanna go through the entire timeline right now, but what I wanna get is a perspective of – I’m assuming you were solving your own problems, that’s why it existed in the first place, and why you released it as open source, because that’s the developer heart you have and whatnot, so it makes sense why you would do that. Did you think that in those moments or earlier on was it a dream that someday you could form a company around this? You mentioned certain indicators or certain evidence… Obviously, you needed to have the evidence, but did you have the foresight to think that evidence would soon present itself that this open source project could potentially be something where you build a business around it?
[00:07:33.14] At that time definitely not. I started using React early 2014; React was open-sourced in mid-2013, and so when I started using it, it was about six months old. It was still very crazy, like “What is this technology doing?” That was most people’s opinion of React… But I started using it, immediately fell in love, and I just had a very strong feeling that this was kind of the future of everything on the web.
When I started working on Gatsby, React was still 10x-20x smaller than it is now… So it was becoming a big deal, but it was still very early in its lifecycle. So I was like, “Okay, I think this is a really important problem that needs solved”, and I had the inkling that some tool in the React for websites sort of space would be a really big deal… But I wasn’t really thinking about it in terms of turning it into a business. That came later, after React had grown a lot, and after Gatsby had grown a lot, and after I had also just thought about the problem space a lot more.
You mentioned you’ve been involved in startups your adult life… Let’s go back a little further to - at least as far as I can go that I’m aware of, which was in the Drupal community, whenever you were doing things around there… I’m reading a quote from you where you said you even attempted to build a startup based around Drupal after college, and I’m assuming that was a fail, just considering where you’re at today, so…
…let’s talk about maybe indicators there, or maybe lessons learned around “You know what, I have better hindsight now, so my evidence actually wasn’t that clear… Maybe I was premature. Maybe I was right on, or it was just a different world.” Tell me more about this.
I mean, I just didn’t really know what I was doing back then. The code wasn’t great, I didn’t understand really anything around marketing, sales… Just a lot of fundamentals of what makes a startup tick, I didn’t really understand. It was very much kind of a blind faith, youthful enthusiasm. I was just like, “Hey, there’s all these obstacles and problems, and I have no idea what I’m doing, but this sounds really cool, so I’m just gonna plunge ahead, and be really determined and run at this thing.” That sustained me for a while, but then eventually I kind of wised up that the mountain I was trying to climb was very steep, and might not lead anywhere interesting. The view at the top might even not have been anything.
Maybe not even the right mountain.
Exactly, yeah. What I was working on was kind of like a social learning – this was [unintelligible 00:10:00.15] times, and… What’s funny is there definitely was something there, as evidenced by Udemy, and Instructure, and a bunch of other companies that actually emerged around the same time, and are very successful still, ten years later… But yeah, I just – I didn’t have the right angle, I didn’t have the right product insights, and I was trying to do it by myself, which was too much; I didn’t do enough customer research… There were all sorts of things that I didn’t understand well enough.
I think that’s one big lesson I’ve learned about startups - you have to get 100 things right. It’s not enough to have like 70 things right, when you’re missing 25 things… Because any one of those 25 things can just blow your plan out of the water. The bar to be successful is super-high. You just have to research and think about things very thoroughly, to not waste time and leave yourself very vulnerable to problems down the road.
Let’s talk about the importance of open source in business in general, and maybe in particular Gatsby Inc. I mean, obviously, it’s important to the project, because hey, it’s open source and built on open source, and all the dependencies; that’s clear based on your dependency tree… But I look at it and I see this perfect storm in your history, of like learnings from Backbone.js, Node, Npm, many people going down this road, and as you’d mentioned, React, having enabled all of this for you to build… And the important point here being the importance of open source and the countless of maintainers, contributors, users of open source being this catalyst for your story of Gatsby to be reality. How does that resonate and translate into you as a founder and CEO taking on investment, building a company around this open core project and business you’re building…? All the countless hours there from so many people being this perfect storm timing? The change of the web, the old way to the new way… Take me to that story.
[00:11:53.14] Gatsby is definitely standing on the shoulders of giants. We see ourselves as a part of a much larger ecosystem, and we want to be playing an important part and be a good member of that ecosystem. I have a strong appreciation for just how much everything we do is all dependent on so many other people, and a lot of what makes companies successful is forming good relationships with all the different parties, other open source vendors, individuals, and just making sure that everyone is aligned on where we’re going and what’s important, and working it all together.
Node being fast and stable and awesome is a huge part of Gatsby being successful. Webpack being fast and awesome and stable is a huge part of Gatsby being successful. Babel, React etc. etc. And part of my goal with Gatsby really is like – I see all these great tools, and I’m like “They could be so much better if there was a really good business model associated, running around all this, because then we could funnel more money into making these tools better, which would then drive the business as well.”
Is that a part of your goal with Gatsby? I mean, obviously, it’s further and farther reaching, because maybe now you’re still improving tech, improving product, or finding traction; you’re still in the finding out what your true product is, so you can actually sell it… There’s some work, there’s some value, there’s some opportunity… I’m assuming you’re still in the proving grounds of the future, but your heart desires to lift up other open source projects that are beneath you in your dependency tree, because hey, like you said, Webpack needs to be fast to enable Gatsby to do what it does, and on down the three…
Right, right. My whole goal with doing startups is – you know, we all live in this world, and we want it to be a nice place to live for ourselves, of course, and for the people we care about, and then future generations, and everyone else in the world, in these kinds of concentric circles, and so forth. So I want to contribute to making the world a better place, and be effective at that, and so, going back to open source - open source is a big part of how the world works. Software is a big part of how the world works, and most of the software that’s around is open source, so helping figure out how to make open source work even better, which will drive a lot of other improvements to the world, is to me kind of a pretty obvious, good thing to work on. That’s why I like open source so much - it’s effective at building great software. And that’s a lot of why I appreciate great businesses - great businesses, particularly great open source businesses, are even more effective at building great software, that lots of people benefit from.
We’re definitely in the earlier proving grounds, like “Does this model even work? Does our product even matter?” etc. etc, but if this thing is wildly successful, what does it look like? …it’s that Gatsby is driving lots of fundamental improvements to web tooling… And we can justify that, because the better all those tooling gets, the better Gatsby gets, which means all our customers using Gatsby become even more happy with Gatsby, so then they pay us more money for services and products, and so forth. So it’s this nice virtuous cycle, where we make people happy, customers love Gatsby, customers love our tools, and so they pay us more, we get more customers, which drive even more improvements to the underlying tools.
Let’s move to earlier this year, May 2018, with the company announcement blog post… You’d mentioned different things in here, and I wanted to kind of break down a couple different ones… Let’s define the line between Gatsby the project and Gatsby the business. Let’s start there, and let’s start to chisel away at different business decisions you’ve had to make. Help me explain the great question, “What is Gatsby?”
[00:15:58.06] The fundamental goal of Gatsby is to improve the developer experience around building sites, and to improve the user experience for visitors to the sites. Those are the two questions we ask ourselves all the time, it’s like “How can we make development smoother? How can we help people who are unfamiliar with different parts of the technology understand errors, or not get blocked by trivial things that we could smooth over? What kind of new abstractions can we create, that provide more powerful primitives to developers trying to solve complex problems around websites or apps?” etc.
Then on the user experience, what is the end user experience around visiting sites? We’re always very concerned around how to make Gatsby sites faster. Ship less code, load faster, lazy-load exactly the right things, so that anytime you want to do something, everything is already ready… So it should just be a very smooth, glitch-free, fast experience for using every Gatsby site that’s created.
And what about the other side of that question, where you’ve got Gatsby – I mean, obviously, you have to define a line, because someone says “Hey, Gatsby”, they could be talking about the company, they could be talking about the open source project, and at some point, potentially, some sort of product that may or may not be announced… So you can help me unravel where name choosing and whatnot - how that starts to define, “Hey, here’s the project, here’s the company, here’s the products.” In the naming scheme you’ve chosen, it could get confusing.
Absolutely, yeah. It’s a very tricky issue for any open source company that’s trying to – yeah, figuring out where that “line” is, it’s difficult… And it really impacts people’s perception of the project, and also just the commercial success of the company.
For us, there’s a few different ways that we look at this. One, I think a really important question is… Open source users – again, it’s kind of a spectrum. There’s tons of people using Gatsby for personal projects, side projects, whatever, where there’s just no commercial aspect to what they’re doing. There’s no money exchanging hands, they’re not making money… It’s just pure play, fun, whatever. So even if we wanted to charge them, they would never pay, because it would never make sense.
Then on the opposite end of the spectrum is people using Gatsby to build out sites where they’re investing, between people cost and tooling cost, millions of dollars. There’s lots of projects like that. We want to target all of our products and everything that we commercialized towards areas where there’s commercial activity, because we think that’s very reasonable. If you’re getting a lot of value from Gatsby, it only makes sense that we should be able to capture some of that value, so that we could then drive more improvements to the project. We look for areas to commercialize where we can provide a lot of value to commercial projects. That’s one way that we think about it.
Another way we think about it is what kind of problems is open source good at solving, versus what kind of problems is open source not good at solving? I think something that’s pretty straightforwardly true, or self-evident just from the last 10-15 years of history is that open source is not as good for running services. Any sort of cloud service - open source can help there. Open source is a lot of the code that’s there, but it often involves very skilled people to build out systems and to maintain those systems in order for them to work.
Open source is at the base of these cloud services, these cloud infrastructure things, but there’s just a lot of skill that’s needed to run and maintain those at a high service level and high performance level, and so forth.
[00:20:13.18] So far, kind of to complete our vision of Gatsby, there just needs to be in a lot of cloud services, which kind of tie everything together, and these are not something that the open source process is very good at solving, and these also are things that tend to be needed by commercial projects. Anyway, that’s kind of where we’re trying to put our commercial efforts into - building out cloud services, which are kind of the glue; it takes the open source aspects and takes other pieces of the ecosystem and builds out everything that a commercial project would want, to have a really great experience using Gatsby.
Right, because… I mean, I can pull down Gatsby, I can put it on Digital Ocean, or Linode, or name your server or cloud - I can do those things myself, freely, because it’s open source. It’s the whole point. But if the value capture, as you had said earlier, of Gatsby Inc. the company - where does that happen at scale, so that we can provide and build out services to make it easier, faster, better, more reliable, whatever…? It doesn’t stop me from being able to do it on my own, but if I graduate or I don’t wanna manage my own stuff or whatever, I can come to you at some point, and you’re turnkey. Is that the direction of the company?
Yeah, exactly. Building a Gatsby site with 100 pages is pretty trivial; a lot of people can do that. But writing a Gatsby site with a million pages - all of a sudden that’s actually really hard to get things working at that scale. So there’s that aspect - how do you actually scale at Gatsby? There’s another aspect, which is kind of the whole control plane, as some people call it, around an open source project. So you’re working on a project, you [unintelligible 00:22:04.16] on GitHub, and off you go. But what if you have 12 people who are all working on the same Gatsby project? And what if you’re adding and removing teammates every so often, and you need to worry about permissions, and you need to worry about who actually can deploy things, stuff like that? Once you get into that area, you need some sort of system around managing access, and deploys etc. That’s another thing that companies are definitely willing to pay for, and is a valuable thing to offer… And again, that’s not something that really makes sense to do in open source.
Again, back to something that people like you and I may see and it’s less normal or visible to, say, folks who aren’t engineers, or software developers - the fact that you’re an engineer and entrepreneur, in one. And I believe that people like you have the ability to see opportunities where others do not see any opportunities, simply because you’re aware of the technical implementations of something, you see the visibility of old ways of web and new ways of web, you put out something to scratch your own itch, and then you discover that “Hey, people don’t actually just wanna do static site pages; they also wanna plug in other services, and this is a really big deal.” You are able to see opportunities where others just see a website. Can you touch on maybe the opportunity [unintelligible 00:23:30.11] just the fact that you’re an engineer, and maybe what engineer founders like yourself see opportunity in? …and maybe even the scale at which – I mean, you took on a seed round of 3.8 million dollars (not a small number), you’ve got great goals in line, but it’s something that came from you being an engineer first, and then having also an entrepreneurial background.
[00:23:55.22] Exactly, yeah. I mean, innovation and invention is something that’s really interesting to me. It’s like, “Why do some people figure things out and others don’t?” People say that creativity… It’s not that people who are creative are special really, it’s just they know a lot of weird things, and then their brains are maybe better – maybe it really is just that they know lots of great things, and then they just connect them, and everyone has that ability.
It’s also very clear – if you look at someone who’s “creative”, they’re often creative in areas that they know a lot about. So it’s like, why does somebody come up with an idea and other people don’t? I think it’s just some combination of knowing the right things and having the right experience, and then having the right incentives to think about it, and then pretty reliable then something will pop up.
I often wonder why isn’t anyone else working on this in the same way… I keep asking myself that question, because it’s like, either we’re crazy and there’s nothing here, or for whatever reason nobody else has thought about this. And yeah, I think a lot of it comes back to the fact that I just have a very unusual combination of things that I have experienced, or things I know that most people don’t.
One of the big ones that you touched on is being an engineer and entrepreneur-minded is fairly unique, because a lot of engineers aren’t interested in entrepreneurship. There’s lots of other engineers that saw that React was big, they knew that it would be a good thing for – React and websites made a lot of sense, but then taking that to the next step, saying “Oh, this could be a business” is kind of a leap that you need be both interested in entrepreneurship and have a good sense of what kind of business models there are, and see the tech trends, and then combine those all together.
Also, I think a lot of why I saw that Gatsby made sense is just my background at Drupal, doing CMS projects in college and after college, and also working at Pantheon, which is kind of a Drupal/Wordpress developer tools and hosting platform, and just having a lot of experience seeing what the world of building websites is like, and knowing a lot of people in that space. It was a combination of all these different things that made it so that I could see something that perhaps other people didn’t.
What I find interesting is how some people are uniquely positioned, while others aren’t, and sometimes often armed with the same information. Why you? …you know what I mean?
Yeah. Do you know the idea maze thing from Chris Dixon?
I’m not familiar.
Basically, what he said is figuring out a startup idea is just like traveling this maze, where you’re just going through all these twisting corners, and ideas, and it really just is a lot of hard thinking to come up with something that’s actually generally useful. And I think a lot of that is just because there are so many people trying to do startups, so there’s a really high bar to meet before you get out of the obvious ideas that a lot of other people are already trying.
A lot of people are like “I’m gonna do entrepreneurship” and then they end up doing – you know, for people in the know it’s very derivative, very obvious stuff, and there’s already five startups doing it. Sometimes I talk to people who are kind of interested in entrepreneurship and they’re coming up with ideas, and just because [unintelligible 00:27:33.14] they’re like “Oh, I wanna do this, I wanna do that”, and I’m like “Oh, here’s five startups that are already doing those things”, and they’re like, “Oh…”
To get beyond the noise and to get into virgin territory requires both a unique set of experiences or knowledge, and also just a lot of hard thinking.
Let’s start with Gatsby the company at this point. We talked earlier about evidences and certain indicators that made you feel a certain way, and earlier in your blog post in May 2018 you mentioned that you started to feel confident that it was time to create a company devoted to bringing the full vision – I’m gonna back and say, in quotes, the “full vision” of Gatsby to fruition. Take us back to there. What were some of the indicators that were pulling you? Obviously, you were doing consulting, and different things… It was commercialized/commercial opportunities, but let’s unravel the onion and maybe even unpack “full vision”, what you meant there.
As I mentioned earlier, our mission is to improve developer experience around building websites, and to improve the user experience of actually visiting, interacting, experiencing the website/app. There are a lot of ways to do that. One way you could think about improving something is you’re like “What are all the workflows or the steps that you take to do something?” Then you analyze those and then you say, “Okay, how can we make them better?” Better can be faster; if you have to fill out this form and it takes two minutes, and you can make it one minute, you’ve made that better.
Twice as fast, it’s better.
Yeah, twice as fast. You can make it more approachable, accessible, so it’s like right now only 10% of people who would want to do some workflow or do something can do it, and then you make it accessible to 50%; then you’ve made it substantially better, because all of a sudden 40% of the people who are interested all of a sudden are able to do it, so you’ve added a lot of value there.
And you can make it more bug-free, so it’s like – you know, you can accomplish something, but it’s very tedious, and people frequently get [unintelligible 00:31:33.16] by irrelevant stuff, so you can just sort of smooth out the workflow, so that it’s less error-prone, and so forth.
Then maybe kind of the final one I think is under-appreciated, but super-valuable - what if you can just eliminate the need to do a workflow or an action altogether? If there’s a step, or even a whole class of things that through some sort of innovation that piece just completely disappears. I think it’s under-appreciated, because people just sort of assume that the state of the world as it is now must always exist. It’s just much harder to imagine a complete – not a complete, but it’s harder to image a part of the world as we see it existing now just disappearing, versus something that concretely exists and then making it better.
[00:32:34.06] This is the classic kind of disruptive versus sustaining innovation, where sustaining is like “Okay, I’ll make the same things we have and make them better”, and disrupting is just completely upend the order in some fundamental way.
Do you feel like you’re in the disruptive category then?
Yeah, that’s what we’re doing.
[laughs] “Yes, Adam…”
Yes, just to be clear… [laughs] I mean, we’re doing both. I brought that up just kind of a background, because part of what we’re doing is taking things that people have always done and just trying to make them better and smoother. But yeah, a big part of what we’re trying to do and why we think what we’re doing is really important and valuable is there’s just things that people do right now that we don’t think you should have to do at all.
One big part is that a lot of building websites is “performance tuning”, where a lot of web frameworks, your sites are kind of slow out of the box, and with sufficient effort and sufficient skill you can make your site fast. Our take is “Why shouldn’t every site just be ridiculously fast out of the box, and then you don’t need to think about performance in kind of some extreme, where you’ve created a very complex page and then you have to worry about lazy-loading stuff, or whatever?”
So yeah, what if we could just eliminate the need to think about performance altogether, and all you have to do is just build something and you can feel confident that it’s going to be really fast?
Another big problem that people run into is running your website. People spend a ton of time setting up servers, setting up databases, monitoring them, responding to outages, responding to traffic spikes, on and on and on, and we’re like “Why shouldn’t scaling and running the website be completely a non-thing? Why shouldn’t it just be automatic?” You can put up a website and it can go from ten people visiting it a day to ten million people visiting it, and there’s just nothing more to do. Because you talk to a lot of people experienced at web launched, and they just dread the launch, because they’ve just experienced – you know, on the biggest marketing day of the year they launch a new thing, and then there’s some bug or whatever, and then the whole thing just isn’t scaling, it’s super-slow, and there’s lots of errors, and whatever.
I can recall talking to Eran Hammer several years back around Black Friday and Walmart…
Oh, yeah, yeah…
Is that an extreme example, or a pretty relevant example?
That’s probably the most extreme example, but… I mean, lots of people have Black Friday type marketing events, and lots of smaller marketing events. We were talking to a company that switched – well, I guess in the process of switching their sites to Gatsby from Drupal, and they said that every time they had a trade show, fairly routinely their sites would fall down… And then since switching to Gatsby, they’ve had zero trouble.
You’re right though… I mean, you shouldn’t have to worry about those things. I guess maybe the disruptive side of you is that there’s several paid services or products that enable scaling. Heroku Dynos, for example, or whatever… Like, you know, rinse and repeat on down the line of hosting infrastructure or cloud infrastructure; a lot of what they have built over the years along the way depends on not having you in the picture.
Someone who is like “Why can’t it just be performing out of the box?”
Right, yeah. We’re not claiming to have invented this idea…
No, I’m not saying you are… I’m just saying you’re the disruption, because you can change it for them.
Yeah, we see ourselves as part of kind of – one way to think of ourselves is we’re like a cloud-native website framework. Or if you think serverless - we’re like serverless websites, where the whole premise of the cloud, and even more so serverless, is that you don’t have to care about the underlying hardware, that computation resources are available on demand, and for the most part scale-free way, so that you can just say “I need resources”, and then as things go up and down, it just transparently kind of horizontally scales itself out to handle the increased traffic, with no intervention on your side.
So our whole premise is that this should be the state of the world for websites. Handling traffic, handling spikes shouldn’t even be something that you have to think about. It should just be baked in.
Performance, scaling… What else?
Another big part is just setting up development environments. It can be pretty tedious with a lot of web frameworks to set things up… Where with a Gatsby site, it’s a Node project, so it has a package.json, and it describes all your site’s dependencies, and you install it, and voila, you’re done. It’s very simple to replicate the production environment locally, without a lot of setup work, which is huge if you’re onboarding new people.
Another thing that we’re trying to eliminate is – a lot of work in web projects is really just getting the data that you want in the shape you want it, to the right place in the website, and you have to often write a lot of custom code to do that. Our contention is that we use GraphQL to query the data, for people to query the data into their project… And our contention is that that’s all you should do.
The whole premise of GraphQL - which we really love, and why we baked it into Gatsby - is that all you should have to do as the front-end engineer is say you want data, and then the data just shows up. We orient everything around that, in that we have all these ways of hooking up to different data sources, so that if you’re the site developer, you’re like “Oh, I need to pull in data from this CMS, or from this job tracking system, or from this e-commerce platform”, and you just add a plugin and then you can immediately start writing GraphQL queries in your React components in order to pull in the data that you need… Which is a pretty dramatic difference than where you’re like “Oh, I need this data, so I’m gonna have to write a bunch of custom code to fetch it and transform it in some way, and push it into the right place in the website.” Which, again, it’s not necessarily – and I think a lot of people forget about eliminating steps, is they’re like “Oh, that’s just a couple hours.” But when you have something you’re trying to do and it goes from – anyway, these couple-hour steps can be real blockers from trying out stuff, or efficiently shipping things. Say it’s four o’clock and you’re like “Oh, I need to do this thing”, and you’re tired and you don’t really have a lot of energy, you don’t wanna dive into something new. But if it’s like a three-minute step, then you can whip out another feature.
[00:39:58.04] Right, if it’s easy to experiment or not a huge time-sink to experiment, it’s a little more likely that you’ll have people taking risks that can be absorbed easier as a startup, especially – you know, maybe some of your customers aren’t exactly in this sort of phase; maybe they’re an enterprise and they’ve got great revenue tracks or whatever, but the point is is that if you enable experimentation at low-risk levels, innovation happens.
Exactly, yeah. If we drive the cost of trying anything down as low as possible, the more we do that - yeah, the more experimentation… I like to call it tinkering, or [unintelligible 00:40:34.22] A lot of building a great product, or a great website, or a great app or whatever, is just you tinker with all the different possibilities; you go down all these different rabbit holes, and then you build this feel for how everything should work, and then eventually you say “Okay, I’ve explored all these different possibilities and I understand all the trade-offs, and voila, this is what we should do, and this is the right setup for everything.” And yeah, if it’s cheap and easy to quickly go down all these rabbit holes, then you build up that mental model, that intuition around how everything works and what are the possibilities much faster, which means that the resulting end product is gonna be much nicer… Whereas if it’s hard to do that, you’ll probably get all waterfally, and just sort of make a decision upfront about how things should work, and then…
You could be wrong.
…you build it – and yeah, you’re probably gonna be wrong in some way or another, but it’s so expensive to build it out that you end up just going with whatever you originally chose, even if it could be suboptimal. So yeah, we’re big believers that enabling that experimentation and tinkering pays huge dividends into the resulting quality of the product that gets produced.
Clearly, you and I can have sort of a person-to-person fairly easy conversation around this when we talk about your full vision, and we started to get here by talking about the blog posts you shared earlier this year, May 2018, whenever you announced the company, and then you also mentioned that you were very lucky to find some great investors, so I’m imagining that part of finding those great investors was sharing a portion (if not more) of the story we’ve just talked through there, which is like “Hey, we can really disrupt and innovate, and reduce to zero, in some cases”, and that became convincing to these investors.
Your seed round was 3.8 million dollars coming out the gate; I’m assuming a part of what we just talked through, the vision, was shared with them… But in your own words, you said we were lucky to find some great investors. Can you talk about what that means? What part is luck, and what part is finding, and what effort was involved on your part to get to that point?
Yeah, so both my co-founder and I - this is our first time raising venture capital. The luck was we found people we knew, friends of friends who offered to help us out. We were introduced to people we didn’t know existed when we started the process, who turned out to be perfect. So whether you call it luck, or whether you call it whatever – at the end we felt grateful, in whatever was the underlying cause for that happening.
I guess what I’m trying to get at is to try to unravel the effort level to find the necessary money to begin to iterate on your confidence that you were given, based on your current circumstances… Because you could have confidence for several months, if not half a year or whatever, and not find money, so I’m curious what your journey was on finding the right kind of money, the right kind of investors… Because it could have been great timing, it could have been whatever, and I’m just kind of curious of the story there. You describe it as luck, and I’m wanting to know more about what’s that really mean.
[00:43:59.02] Sure. Raising money - there’s so much that goes into it, but it generally boils down to “Is there a good fit between the founders and the market?” Do the founders have some fundamental edge on delivering something really valuable to the market? Basically, our fundraising was like “Hey, look at how awesome Gatsby is. There’s tons of people using it”, and then we said “Hey, look, it’s aligned with all these really big trends around serverless, cloud computing in general, React, all the changes in the front-end tooling space.” Then we explained “This is how we’re gonna make money”, and then they were like, “Okay, you clearly know a lot about this market, you’ve already proven that you can build something that people find really valuable, and all the trend lines are just going up and up and up. This is already a big space, and it’s gonna be a far larger space in the future, and your model for how you’re gonna make money sounds plausible, so let’s do it.”
The main thing is investors don’t need to have a high risk of certainty around an outcome; the main thing that they want to know is “Can this be really big?” 5% odds are just fine if the actual eventual outcome could be quite large.
I was kind of joking to a friend when we were fundraising… It’s like, a lot of fundraising is just pretending to be a megalomaniac and telling everyone you’re gonna take over the world, and sort of believe it, and then that work. Investors get really frustrated when people are like, “Oh yeah, we have the best plan. We’re making tons of money, this is so awesome”, and then when they push at it a bit, it’s like the total addressable market is 500 million, or something like that. They’re like, “Well, okay, that’s fine, but that just doesn’t match the VC model, where they want to invest at a low price in something that’s highly risky, but can turn into something huge, even if there’s only a 5% odds of that happening.
Right. Is this your first time raising funds?
It is, yeah.
How did you know that you can trust the people you’re working with, and then how stock gets issued, or the finances get handed over? Were you scared, or – describe your personal demeanor in the process.
Yeah, I mean – I’ve been to Silicon Valley for almost eight years, and I’ve had friends who raised money, so it wasn’t a totally foreign thing. I’d never been directly involved with it, but it wasn’t a foreign thing. And I’d explored the idea before in the past, and I’ve read quite a bit about it… So the whole process was reasonably familiar, at least the outline. Then, when we started getting really serious about raising money – I mean, we reached out to different people that we knew, and met with them, and they gave us a deeper crash course on how everything was gonna happen… But also we found just that the VCs themselves are very used to explaining how stuff works, because they know that there’s a lot of – I mean, it’s their thing, every day and all day, whatever, but they’re very used to first-time raising money entrepreneurs to come in and not know the process… So they’re very used to explaining how things work.
Yeah, they get you out of the fear, uncertainty and doubt pretty easily. They kind of anticipate it, so they’ve got their – for lack of a better pitch, like “Hey, chill out. We’ve got your back here, and here’s how we’ve got your back.”
Yeah, yeah. And VC-entrepreneur relationships can be adversarial, but that tends to be when things go south, later down the stage. You never take money unless you fundamentally trust the investor in some way, because you’re putting a lot into their hands, and they’re putting a lot into your hands… Both sides are highly incentivized to be frank and open with each other about why they do what they do, and whatever. And we knew that, so we just asked a bunch of questions and they were very helpful. Then we had our advisors who had kind of been in the trenches, doing things, who helped explain a bunch of stuff as well.
[00:48:14.19] My emotional demeanor when we first started - I was just nervous, like “Can we even raise money at all?” We were both plunging in just for the first time. We were pretty sure this would work, but we were like “Will this really work?” But then, within the first week and a half, we had two offers, so we were like “Okay, we can definitely raise money.” Then at that point it was more “Who do we wanna work with and how do we figure that out?” We just took a bunch of meetings with different VCs, and asked lots of questions, and asked our advisors, and then over time we kind of understood our range of options, and what it would be like to work with them, and then from that we made our ultimate choice.
What were your primary concerns and maybe what were some questions you asked to ensure that your – maybe not your visions align, but your interests in partnering… Because that’s what a VC relationship is - it’s a partnership based on funding, of course, but sometimes they even bring advisory roles to the pitcher, or connections… What were some of your concerns and questions back to them that it was like “I’ve gotta ask them this”, or things that got answered that you were like “Okay, you’re a perfect fit for us.”
Yeah, I think most misalignments between entrepreneurs and VCs is the entrepreneur doesn’t really want to build like a rocket ship company, that’s gonna grow very quickly for a very long time… Because that’s all VCs really want. I mean, that’s all VCs want to invest in.
I think a lot of entrepreneurs say “Oh, wow, I have this cool idea”, but either they haven’t thought through it enough to understand whether it could be a very scalable startup, or secretly they don’t want to be part of that journey, which can be kind of tumultuous. So maybe they raise money, but then they don’t act in a way that will help their company grow very quickly, for a long time… So because they’re not acting like that, then the VCs get frustrated and start to be more micro-managing and pushy and whatever, and then you have these weird conversations where – I don’t know, the entrepreneur is just not admitting their real motivations for doing things, or their real fears or whatever, so the relationship goes south, because then the VC is like “Oh crap, I invested in someone who isn’t interested in building out the kind of company that I was expecting that they were gonna build out.”
Right, so having clarity, and maybe even some clear expectations of “Hey, I need money, I need roughly this amount for these reasons, and here’s our pitch”, and then they’re like “Well, I want an entrepreneur to invest in, or an opportunity to invest in that aligns with this kind of growth model, or this kind of speed, or efficiency or whatever” and you’re like, “Okay, that works. We can do that. That’s what we plan to do.”
Yeah, exactly. I mean, everyone of course wants millions of dollars that they can do whatever they want with… Because that’s basically what you get - you get millions of dollars that you can do whatever you want with. Who wouldn’t want that…? But if you go into it not fully realizing the responsibility that you’re taking on, which is to deliver a really great return on that money, and you just kind of want millions of dollars to throw great parties and look cool to your friends, or whatever…
That’s what I wanted to get at - what are your motivations with the money? What are you gonna do with it? Because we talked earlier about your history - engineer, entrepreneur - some of the reasons and indicators that gave you confidence and evidence to move forward, and it was like “Okay, great”, now you can move forward, turning this into a company… Are you capable, do you want to, and are you willing to change – I assume you have great answers for this, but like, this is what I think about… Like, do you want to, are you willing to do what’s necessary to take on venture capital? I think that you have to be a certain type of person in a leadership role, in a startup, to say yes to something like that, and be responsible, to be able to follow through.
[00:52:18.05] Exactly, yeah. Both my co-founder (Sam) and I thought a lot about this and talked about this. We were definitely up for the challenge of a venture-backed company, and we knew that Gatsby had the potential to be the type of company that somebody would invest in… So that part was very straightforward. And VCs are definitely trying to sniff out people who aren’t serious about that…
It’s a part of every process, right?
Yeah, so we passed that pretty easily. So then it was just like – the uncertainty we had was like “Well, we think this model makes sense, but maybe we’re missing something, or whatever”, so we just weren’t sure at first if investors would believe our model… Because it’s kind of novel in some ways. But yeah, as it turned out, we found investors who did have experience with other companies who were doing similar sorts of stuff to us, so they immediately saw the potential, and things went pretty smoothly once we found them.
Let’s talk about where you’re at now then. Obviously, you took the money, you forged a relationship, you have a great partnership, I’m assuming, you’re several months down the line - that was May when things kicked off, or at least the announcement… I’m sure that it was several months prior to that that you were discussing and reviewing and deliberating and whatnot to go that route… Now we’re barely a week away from October the same year. What’s happened since May, or the announcements, and stuff like that? Or not so much what’s happened, but where are you at in terms of iterating on those promises?
Yeah, yeah. So for an open source business to work, you have to have – like, the commercial business aspect of it is entirely dependent on the open source part being really amazing, so our first priority after taking money was to basically scale up the open source sides of things, to just really accelerate the improvements there.
Our first four hires were – we hired a UX researcher; she just talks to Gatsby users all day long, and then translates that into improvements both to our open source product, and then also the tooling infrastructure in it. So we have gatsbyjs.org with a bunch of tools on it, and also stuff within GitHub, and so forth.
Then we hired a designer, who does both designs and does a lot of front-end engineering around all these different kind of projects, like the .org site, and also our .com stuff, commercial stuff.
Then we hired two – there’s two very active people in the community who were helping out a lot, answering questions, reviewing PRs, and also writing quite a few PRs, and we were just super-impressed with their work, and so we reached out to them and said “Hey, would you like to work on Gatsby full-time?” And they were pretty flabbergasted, because they didn’t even know that would be a possibility, because they didn’t know we were raising money, and so forth. After the initial shock subsided a bit, they were like, “Um, yeah. That’d be amazing.” They just really liked the project, and liked the vision of it… Anyways. They’ve been fantastic, and have freed up a lot of my time, and also just really pushed the project forward a ton. Then we hired a developer relations person, Jason, who you talked to previously on here…
[00:56:08.08] If you think about open source, if you think about any product, it’s like “Where you create value?” You create value from having a great product, but if nobody uses that product, then the value created is kind of the potential value that can be created per person, and then the number of people using it. So it’s kind of like “X times Y equals the actual total value being created.” So we wanted to make the actual product a lot better, but we also needed to evangelize it, so that more people would understand – it’s like, “Oh, Gatsby exists, and Gatsby is actually a great fit for all these problems that I have”, and then start using it. Jason has been fantastic about doing that.
That was our initial hiring approach - to build out the team around the open source project, and driving that forward in a bunch of really critical ways. That has been amazing, just to see them really start to own all these different areas and drive a lot of really awesome improvements there.
Then our next focus has been around just our cloud infrastructure and commercial products that we’re gonna be building. We’ve made several hires there, both cloud infrastructure engineers, and then also we have a marketing and sales hire that we’ve made, who are starting to build those types of things.
Did you get any advice with this strategy? Is this something that you and Sam came up with, or were your venture capital folks involved in this, or was it just you and Sam and team doing and building?
I mean, this is…
Sure, there’s a mix, too. This is a reasonably standard OSS growth strategy. Like I was saying, you have to have tons and tons of open source adoption before you can have a great business. That’s why I didn’t even think about starting a company until OSS growth was quite a ways along, because I needed to validate that it could grow large enough that business actually made sense. And then even after we raised money, the focus was “Okay, let’s set the foundation, so that Gatsby can grow to even way better than it is now”, and get the right people in place to shepherd that growth. Then once that was sailing along, then we started moving focus to worrying about the commercial side of things.
It sounds like you think it’s common knowledge, and I would say potentially… You know, what you’re doing and what you’ve done. I would say that the reason why I think this conversation on this podcast makes sense is that it’s not common knowledge. To you it seems, because maybe you’re so close to it, and you’re in it every day, and that’s all you think about, it’s common knowledge. But I would think that the types of hires you made, the strategy behind them, and the reasons why you think they’ll work, your confidence, is not extremely common.
Yeah, I’m not gonna disagree with that. It’s common to the small group of people that think about this stuff, and have experience with this stuff.
The thing is, there are blog posts that line all these things out, so… Yeah, and you can look at – there’s been a number of very successful open source companies; Elastic is a more recent one, MuleSoft… So you can also just look at what other – MongoDB is another great one… Anyways, you can look at what these companies have done and learn a lot from these models. But yeah, I agree, it’s not common knowledge in the sense that lots of people know it, but it’s common in the sense that the future is unevenly distributed common, where if you happen to be in the pocket of open source, then it is common.
[01:00:08.21] Right. You mentioned numbers… Late in the game, in this conversation at least, to mention numbers, but just since you’ve mentioned the growth, which we haven’t talked about really at all… In the same blog post you mentioned Gatsby is used by tens of thousands of developers and orgs, and is downloaded nearly half a million times per month. This was back in 2018, so I’m assuming those numbers have maybe doubled since then, or at least some x? What’s the x on those numbers?
Kind of across the board pretty much everything is growing - website traffic, downloads, stars etc. They’re growing like 15%-20% a month.
Maybe personally, how are you transitioning? Because a lot of this conversation was in the thick of product and strategy, and parts of it somewhat technical… Obviously, you’re an engineer-entrepreneur - how much do you struggle with teetering that line of engineer/CEO and your shift and change in role?
It can be hard. I guess the hard part is just knowing that you’re really good at certain things, but knowing that your role demands you to do new things, and then kind of taking the leap of faith that you can transition smoothly, and that you’ll be able to do things… But also that you’ll retain the trust and confidence of people around you as you’re learning those new skills, and so forth.
Yeah, it is very much a leap of faith, jumping from something very comfortable into a new task, learning it fast enough to feel competent yourself, but then also know that you’re providing good help and fulfilling the needs of that role to the other people around you.
And what do you think the number one thing is that you personally are getting right in that transition, and how is it affecting the product and company?
I think a lot of it is just accepting that I’m not gonna be good at everything, and that I can’t do everything… Because I think that was the hardest thing for me at first - I was just kind of like a one-man band in the open source days… So switching from that, where I was writing most of the PRs, reviewing all the PRs, writing all the blog posts, writing all the docs, or most of the docs, whatever - at least reviewing all the docs that came in - to hiring and stepping back from all these pieces… Yeah, it’s just like letting go of stuff, and trusting people who are now owning those roles to do a great job.
So yeah, just knowing that you’re really good at something, and then hiring somebody else to do those things, and then using that freed up time to then learn new roles that are needed as the project grows.
I think it’s also hard just to – I have to understand myself, my motivations, and sometimes it’s like “Okay, maybe you should leap into that new role. Maybe you should just hire somebody.”
Let’s talk about that. I’m curious if you’re a fan of doing something before hiring somebody to do something… Or is it sort of like sometimes, and it depends?
Yeah, I’m definitely on the “it depends.” We hired a veteran sales leader. We obviously needed sales help, and we obviously didn’t know what we were doing enough to actually do it at the scale and level that we needed, so hiring someone who has proven expertise and skills in the area who we trust is just kind of an obvious thing to do.
[01:04:01.04] There’s kind of like well-understood roles and well-understood needs, and actually trying to do those jobs yourself could just be a waste of time, because it’s like, “Well, that obviously needs to happen.” Where it’s more ambiguous, I think - you don’t really know what you need per se; then I think actually jumping into that and trying it out and figuring it out… I think that’s an important part of the process, because you need to actually understand that role.
Yeah, if you have a hard time writing the job post for somebody and you’re not clear what you could even instruct them to do, that’s probably something where you should camp out in a little bit further, so you can get some more clarity on what the job actually is… Because you can’t really lead them well in their mission, which is clarity and expectations, if you don’t understand. But if the role is clear and there are expectations that you can firm up, it’s a little easier to hire somebody, versus feeling like you need to do everything first before you hire somebody to do it.
And plus, how do you scale fast if you’ve gotta do all the jobs, right?
Exactly, yeah. So maybe do the job before you hire the job is not the right way of thinking about it. It’s like, understand the job, and then hire for that job. It could mean that you actually do the job to understand the job, or it could just mean that you talk to a bunch of people.
My co-founder, Sam, he leads kind of the enterprise side of things… So before we hired the sales leader, we first spent a lot of time thinking about “How are we gonna make sales? What kind of sales are we gonna be doing?” Is it gonna be predominantly large enterprise deals, or is it gonna be mostly self-serve, with the occasional upsell to a small(ish) 20k or whatever annual deal, or something, with some support or whatever thrown in. Because those two types of models - and there’s lots of other models of course, but anyways, those two types of models are very different, and then the sales organization is also very different, and the kind of leadership that you need and the systems that you need are also very different, so… We had to resolve that fundamental question ourselves, before we could even think about hiring somebody.
Yeah, you had to figure out who your customer was. Ideal customer…
Yeah, exactly. We had to figure out a lot of things before we could even start thinking about the salesperson, but then once we understood that, then Sam just went and talked to a bunch of candidates and to other people just in general about “Hey, what kind of person do we want?” and then that led to a lot of further refining of our model of what an ideal candidate would be. Then, after we got through all that, then we got some introductions to people, and we interviewed them, and the person we ended up hiring, Jim Ettig - he worked at New Relic, he worked at Loggly, he worked at several other developer companies selling similar types of products, and is very experienced setting up the systems and sales teams to do kind of like… You know, we ended up like, okay, we’re gonna be predominantly high-volume, low transaction cost kind of sales.
Anyway, so we didn’t necessarily need to do any of the work that he is now doing, but we needed to understand the job we needed done before we could hire that person.
That makes a lot of sense. That’s the hard line - one, how do you get the clarity, how do you understand, how do you know? Do you do some of the job yourself for a while? How do you discover certain things? And I think sales in particular is unique, because you’ve gotta understand what you are selling, and then two, who you can sell it to, and then three, how you can inherit repeat buys, or good customer experiences, or renewals, and all these different things that are all – unless you’ve done any sales whatsoever, which I’m assuming when you’re talking about VC funding, you weren’t quite there yet; you were beginning that process, you were still learning what you were building and who you were selling it to and all those things. Even probably to this day you’re still iterating on some level of what you are selling and who you’re selling it to, and why even…
…which I find very fascinating. Let’s close with – and I don’t even know if you have anything, but I’m gonna ask you… We didn’t pre-rehearse any of this stuff, so I’m just curious what’s on the horizon for you? What’s unknown to anyone looking at Gatsby either as an open source user, or somebody who’s like “I can really see after hearing this conversation, checking this technology out, and what you’re doing, and potentially being a customer”, what is on the horizon for you all that no one knows about, that you can either completely spill the beans on, or just tease? What’s upcoming?
Yeah, so we’re really excited about the cloud tools that we’re building, because – I guess you can think about the category we’re in, of static sites, where you produce a site, and then it’s like files [unintelligible 01:09:06.27] It has some fundamental limitations that make it basically unsuitable for all except a small percentage of sites. Right now, the percentage of the web that’s running on the model that we’re working in is half a percent, three-quarters of a percent of websites.
What we’re working on is going to make Gatsby usable for 20x, 30x, 40x more websites, so we’re really excited to make that happen. The fundamental limitation is just around how fast you can build sites, because there’s a build step; so you make a change, and then you have to rebuild the site before it goes live… And our goal is to eliminate the build step altogether. I mean, it’ll still happen, but it will be so fast that essentially it won’t even feel like there’s a build process. So we’re working on making Gatsby incremental builds be very quick, like sub five seconds.
It’s a big R&D project, and there’s a lot of work to make it happen, but we’re very confident that we can do it, and once we make that happen, things are gonna get really crazy.
Very, very exciting. Alright, Kyle, thank you so much for joining me today. I really appreciate you sharing so much. It’s been awesome to talk to you.
Likewise. I’m really glad we had this conversation as well. It’s been really fun.
Our transcripts are open source on GitHub. Improvements are welcome. 💚