The below is a full transcript of a Podcast episode I recently recorded with Joseph Jacks, the founder and general partner at OSS Capital. You can listen to the podcast on Youtube, Apple Podcast, Stitcher or wherever you get your podcasts.
Hello and welcome to another episode of Sand Hill Road, the show where I talk about early stage startups, venture capital and technology. And today I’m really excited to be joined by Joseph Jacks, who’s the founder and general partner of OSS capital, which is a venture capital firm, which has a thesis specifically around commercial open source software companies. And unless you’ve been living under a rock, you must know that commercial open source companies are on fire. Last year, they’ve been killing it, with four billion dollar exits. Github with their USD 7.5bn exit to Microsoft, MuleSoft exiting (for a second time) to Salesforce, Suse being acquired for USD 2.5bn by EQT, Magento being acquired by Adobe for USD 1.7bn. So this space has really been on fire. But interestingly enough, until now, you didn’t have a venture capital firm, which was really specifically focusing on open source software projects. Obviously, you have traditional VC funds, like a16z, Benchmark, Amplify, Redpoint or Index, investing heavily in this space. But until now, you didn’t have a dedicated firm that was exclusively investing in commercial open source software companies. And that’s why I’m so excited to be joined by Joseph Jacks today. So let’s dig right in.
Introducing Joseph Jacks
We’re here with Joseph Jacks, who’s the founder and general partner at OSS capital, which is a venture firm with a specific focus on commercial open source software, COSS. And maybe let’s take a step back. And let’s have Joseph give us a little bit of his hero’s journey. You are now deeply involved in the open source community, how did you get into this? And how do you get from an operator-founder to the venture capital side, maybe walk us a little bit through the steps towards towards entering the wonderful world of venture capital?
I would say accidentally more than intentionally. So I actually originally started my career in sales, sales-engineering and working at different technology companies. One of my very early jobs about 10 years ago, I was working for this French open source, middleware, data integration vendor called Talent, which eventually went public and I worked there for a couple of years. That’s kind of what got me interested in this open source world. It was a very large, successful enterprise software company that built a business around an open source, data integration middleware tool. I worked for a couple of fully proprietary software companies after that, but the last five or six years, I’ve kind of gotten lucky enough to stumble into learning about Kubernetes. Early on, I was an early consultant at Mesosphere, the the company behind Apache Mesos. And when Kubernetes came out, it was very clearly an exciting project on lots of levels. And I was I was lucky enough to help start and build the sort of the first Kubernetes startup company. As part of that I started this conference behind Kubernetes, called KubCon, just kind of the Kubernetes conference and super lucky to meet lots of really amazing people through that experience. The CNCF, which is the cloud native computing foundation formed as the sort of foundation for Kubernetes, once Kubernetes was donated to that foundation, it’s part of the Linux Foundation. And KubCon was also donated to that, because Kismatic was acquired by Apprenda in New York. Apprenda is kind of like Heroku Platform-as-a-Service, originally focused on net Windows Server applications. And they were heavily kind of shifting the company towards Kubernetes. So I was part of that and helped help drive a lot of that stuff there. After that, I worked as a EIR at a storage company actually up in Seattle called Quantum and helped with the early CEPH, which is a sort of distributed object block file storage system, widely used open source project framework. And that project was called Rook and Rook eventually formed the basis for a company called Unbound that I helped, that I helped build and helped with a bit. And then after that, I actually started working on another company, and OSS capital sort of evolved accidentally, for lack of a better story and framing. I’ve always been intellectually obsessed and curious about the intersection of open source software and commercialization dynamics and sort of business building product building, since kind of talent days. But the last, you know, you know, naturally, the last five or six years has been like very intensely focused on this. But I sort of came to the conclusion that primarily because of all of the the sort of economic outcome activity through 2018 IPO, m&a and private equity outcomes, you know, through the rest of 2018, sort of totaling around $70 billion, half of that was RedHat, that the tides were sort of shifting and turning in a big way, in terms of commercial opens source software, companies actually being a real understood category. And then in addition to that, the venture landscape, obviously, having invested heavily in this, but not really, with a lot of clarity, but more tangentially sort of opportunistically kind of gave rise to the opportunity to build something that was exclusively focused on this. And no one was really doing it. And I thought it was a pretty obvious thing that should be, you know, that should be tried. And so I started scratching it down a little bit, I started actually blogging about some of the observations from this, sort of like S&P 500 inspired index of large and successful commercial open source software companies of the last 10 years that have reached considerable revenues, the hundred million dollar thresholds. And the index has kind of grown from, you know, three or four companies about six years ago to now it’s 40 companies, and it’s sort of been growing pretty rapidly every year.
Defining commercial open source software
So let’s get a little bit into this Index, the COSS Index. I’ve checked it out and it’s really interesting. As you said, it was at only a handful companies five years ago and now it’s at 40. So the criteria for choosing these companies is that they are all in the “commercial open source software space”. And so let’s maybe take a step back and look at the definition of commercial open source software first. I mean there’s a lot of questions that come up. Some might argue that even Google is a COSS company because they’re relying heavily on Linux, or Amazon because they’re basically repackaging all this open source software, like Redis with Amazon ElastiCache, or Hadoop and MySQL. Would they also fallen under the COSS category that you’ve established or what’s really like the definition of COSS?
I’ll take a step back and unpack the definition that sort of I’ve been proposing and talking about for a long time, behind commercial open source software company, the sort of COSS acronym and sort of what I view as a fundamentally new different kind of category that that hasn’t really been studied and treated with scrutiny and clarity, in the past. What I would say about the definition for COSS is pretty simple: if the company fundamentally exists, because of the open source software project existing then the company is definitionally a commercial open source software company. And this is independent of the business model, independent of the technology. It’s independent of, you know, license model everything. So, for example, MapR, HortonWorks, Cloudera, as companies would not exist without Hadoop. GitHub, Gitlab would not exist without Git. RedHat would not exist without Linux. Hashi Corp would not exist without Vagrant initially, but over time, they created a few projects. Hashi Corp is an interesting example because Initially, the match is typically one to one. So there’s one project and one company, potentially many companies. So it’s sort of like an potentially N-to-N, but most commonly, 1-to-N where there’s one project and many companies, but it gets a little bit more multi dimensional there in some cases. So a COSS company fundamentally exists because of given open source software project exists. And over time, the company builds a stack of products and differentiated services and hosted offerings or IP, or what have you around that open source software project and derives business value from that in terms of commercial services, and revenue and profit, and product and so on.
It’s interesting, because if you think about, you know, fundamentally proprietary software companies like Google being built on Linux. There you of course the company exists not because of Linux, but it was built around it’s own propriertary internal logic, obviously the PageRank initially, and its seven different products, have been been built around that with their own internal logic and their own different business models. So I think, it makes a lot of sense that you say, instead of focusing on the business model, you are focusing more on on the existence of a core open source project and a community.
So I will unpack it a little bit more, and I do also think it is independent of community. So sort of respond to the examples you listed, Facebook, Google, and so on, I don’t categorize those companies as commercial open source software companies, because fundamentally, the core runtime core business logic, core essence of those companies in terms of technology is proprietary. And with a commercial open source software company, the core runtime, the core essence, the core business logic, the core sort of value creation mechanism for the building block of that company, is an open source project that is licensed as such. So you know, license as Apache to or MIT or GPL 3 and what have you. This can quickly segway into a licensing discussion, but I think the licensing discussion is a little redundant. One simple way of thinking about this, and this is kind of something that that sort of occurred to me recently, which is probably worthwhile unpacking separately. But with commercial open source software companies, the software that creates the vast majority of the value for that company is decoupled from the company. So the software is decoupled from the company. With a proprietary software company or proprietary technology company. And you can include Facebook and Google in that in that category, and probably 99% of all technology companies in that category, the software is tightly coupled to the company, instead of being decoupled it’s tightly coupled to the company. So that abstraction of the software that creates the vast majority of the value, either being tightly coupled to the company, or decoupled from the company is another way of thinking logically about the difference between a commercial open source software company, and a fundamentally proprietary software company or technology company.
Reciprocity of contributions in open source software
So let’s drill in a little bit in into contributions made to open source projects. And so one thing I always wonder about is that in this commercial open source software model, when you look at the Github repositories, you often see that the main contributors are people working at the COSS companies. Of course there are examples like Segment or JFrog, which were sort of built around specific Git repository that got tranction. And then of course, in the later stage of the projects, you often have hundreds of contributors. And this is getting us a little bit into this licensing discussion where we have recently seen the controversy around the Commons Clause, which was penned by your portfolio partner, Heather Meeker. And there the discussion was basically that AWS was repackaging Redis, and that Amazon wasn’t really making a lot of contributions to the open source project. So what’s your take on this, also when you look at potential deals? Do you look at different Git repositories, the number of commits made, by which parties, and is this a core part of your due diligence? And how do you view this issue of, reciprocal value contribtions in open source projects in general?
This is a really broad set of things, but I’ll try to answer what I understand to be the essence of your question. So, we definitely value a relationship between the company and the open source project, as one where the company is regarded as a center of gravity in terms of influence, you know, valuable contribution, trusted with the roadmap, engineering, quality, and so on. And there’s lots of dimensions to that. That does not mean that the company is sort of monopolisticly controlling the entire open source project. For some projects, where there is no single company able to really have a significant degree of influence or you could say control or centralization, around some important dimensions, it is far less likely that a given company could actually successfully commercialize that project. There are a couple exceptions to that. So DataStacks founders, the company behind Apache Cassandra, were just really active community participants in the Cassandra project early on, they helped people understand the technology and learn about the project and adopted enterprise environments is this new kind of NoSQL database. So it’s complicated technology to run in production and so on. They were actually able to build so much goodwill and an influence and credibility in that Cassandra community without actually forming the company with any of the creators of Cassandra or the authors of Cassandra, to actually build a large successful business over time. So there are few exceptions to that. It’s not super common, but there are some exceptions, Fastly is actually another example of a recent IPO, where the sort of the creators of Varnish, which is the project that Fastly is fundamentally built on, do not actually work for Fastly, or were not part of the founding founding team of Fastly. However, fastly came together by building sort of a differentiated product building on top of Varnish, and, you know, basically going going to market with with with a bit of a different sort of more of an indirect commercialization approach. They actually built a differentiated CDN, content delivery network based application and a variety of other things on top of Varnish, and they compete in existing markets against players like Akamai and CloudFront and, and also have new products. So but these relationships get a little bit context specific. In general, though, I think you’re right, we we definitely pay very close attention to the relationship between the company and the project on lots of dimensions.
Yeah, so so maybe let’s, let’s look at the Redis example a little bit more. We had Redis Labs basically being the monetization layer. And I think you mentioned this tweet by Doug Cutting, where he said basically that it’s insane to think that the contributions made to the project will always equal to the value derived from the project. And in the Redis example, we had this situation with Amazon. And the question came up in the community, whether or not to basically limit the commercial use of the software. So basically you have a three-party-scenario where you have: the open source project, then the monetization company, and then a third player cloud provider coming in. So what’s your thinking around this tripartite dynamic, if you like?
That’s interesting. Tripartite, I haven’t actually thought of framing it that way. I would say it’s more of a multipartite, with n degrees of complexity instead of just three. So going back to what I was mentioning, around the decoupling of the software from the company, so you know, now you have these two independent variables. When you decouple the project from the company, and the project is the source of the vast majority of value creation exponentially more than the company can create independent of the project, because it just, you know, the company, the shell, what occurs is, is this sort of meritocratic capitalism phenomena. And this is not a new idea, and I’m not a, you know, a sort of deeply studied economist. But if you were to frame it in terms of comparatively, describing the traditional, fundamentally proprietary technology software company model as being one of monopolistic capitalism, and with commercial open source software, that model being one of meritocratic capitalism. When you decouple the software from the company, you have many constituents capturing value and commercializing around that core project, what is unique and nuanced and interesting. And worth noting about open source software, though, when you have a successful project and ecosystem is that there is a very, uneven distribution between the value capture constituents around the project. And, you know, those value capture constituents are multifaceted as well. So like, for example, the developer ecosystem that uses an open source project is actually capturing value, given that they increase their skill level and their job transfer ability, and their market ability. And you know, Kubernetes is one of the hottest, most valued, you know, deep technical skills now and infrastructure software and DevOps and systems engineering. So if you’re a deeply skilled engineering resource that is super knowledgeable and running Kubernetes, and contributing to it, and so on, like your marketability in your job, your prospects go up by probably a factor of three, or four or five or something, I don’t know. But you know, this is a super big supply demand problem, right now, the Kubernetes ecosystem, everyone’s hiring Kubernetes experts, and there aren’t enough of them. So that’s one constituent that captures the value around an open source project that isn’t a commercial entity. And that’s not something that traditional economics measures, it traditionally only measures commercial entities or shareholders. And I think that’s a big flaw in the system. Now, with the commercial entities that do capture value around the open source project, like cloud providers, commercial vendors, product companies, companies and industry, like, you know, banks, retail, pharma insurance, and so on, those entities or constituents capture value in different ways that actually serve to increase the overall usefulness and effect of, of sort of meritocratic capitalism across all the other constituents, not uniformly but sort of, in a rising tide kind of way. And so what I guess you could sort of look at with commercial open source software, from an economics lens is that it should actually almost be encouraged. In fact, I very strongly believe that it should be encouraged and promoted, and explicitly recognized that the fundamental economic model, because it is one of meritocratic capitalism actually serves to benefit all value capture constituents, when that model is embraced, because obviously, if you build a better differentiated, more compelling offering, you are in a position to capture, you know, a meaningful enough percentage of the value created of that project to actually build a huge business. So I’ll give you an example of, you know, one case where, say ElasticSearch is a project is used by millions of people, the company behind ElasticSearch is five plus billion dollar publicly traded software company or commercial open source software companies, as I would call them. And you also have many cloud providers taking Elastic Search, and commercializing that code base as a service, because the project is open source. That’s actually a good thing in my mind, because it serves to increase the credibility of the project through the distribution channel, and, you know, sort of customer base and credibility of cloud provider platforms, which are very significant and, you know, have benefited from the, you know, 10s of billions in infrastructure investment and innovation that cloud providers have brought on and delivered to the world over the last decade. And it also, it also serves to increase the level of quality in the code base itself, the project, even if cloud providers or these large, vendors don’t explicitly contribute back to the project proportional to the value that they capture, which goes back to the Doug cutting tweet that you mentioned, which, you know, I sort of paraphrase a lot. But Doug says that it’s insane, to expect value contribution back to a project proportional to value received or value captured. Because they’re two dramatically different dimensions of the value. value creation, often dwarfs value capture, and in most cases, 99% of the time, in my, in my opinion, for successful projects, value creation is orders of magnitude more than value capture, for really any constituent except perhaps for the entire developer, stratosphere and ecosystem. If you look at commercial constituents that capture value, I would say that probably bound by maybe very low single digit percentage points of value created. So you could have a commercial open source software vendor capture 1 or 2% of the value and become a 10, 30, 50 billion dollar company and a cloud provider capture 1 or 2% of the value and build a 10 plus billion dollar service, both those things can happen harmoniously. Both those things serve to drive the project ecosystem, that community forward. And both businesses can actually build large, profitable platforms that can sustain their endeavors.
Open source business models
So I think it makes a lot of sense to think about, and I think it’s something that’s been sort of lost in this whole Redis discussion that having the cloud providers involved is actually a way of getting distribution for these open source software projects. So let’s talk a little bit about the different business models, you distinguish between a number of different models: the open core model, packing, SLA and so forth. Maybe just talk a little bit about how you think about different business models, there’s a lot of variants of course, but the kind of categories that you that you personally use as your mental models when thinking about these projects.
Sure, yeah. So there are there are a handful of business models that costs commercial open source software companies implement. And before I actually dig into those different business models, that it’s actually worthwhile mentioning, the different types of models and the sort of like differences between high level models, overall independent of business models, so if you could start at the top level with the company category, the sort of category of company, which is the commercial open source software company, within that you have business models. And then with, you know, in addition to that, you have licensed models and industry that you have distribution models. So these are all like independent of each other. One of the one of the things that I think we see a lot of industry is kind of conflation of a lot of those models. And I think that in most cases can kind of be a mistake. So people sort of think about classifying the company as its own unique, sufficiently complex thing to sort of assess. And separating out a definition for that.That serves to sort of more clearly understand business models within that type of company. And then, you know, license models and distribution models. So for commercial open source software companies, the most successful to date, business model, that large scale companies that we’ve been tracking and assess use and implement, in very, very dimensions is this open core approach. And you can think of open core as a business model, but also sort of a way of constructing, constructing a product and architect being a company and so on, it’s more than just the sort of way the company generates revenue is has lots of different implications. Another business model that commercial open source software companies implement is one of just pure support services and training and consulting. And this basically, is less common among very large, successful commercial open source software companies. So what I mean by large successful ones is sort of north of 100 million revenue in the history of commercial open source software companies, in the last 15 years or so, or really, maybe even since Red Hat, late 90ies, there have only been say, like, I want to say like five at most commercial open source software companies that exceeded 100 million revenue, with a pure support-services-training consulting kind of approach, which, in other words, means no proprietary IP, no proprietary technology, in terms of source code, and so on. Open core, is essentially, the fusion or the integration of the open source project alongside with around underneath a proprietary product of some kind, or proprietary offerings or hosted service, what have you. And by proprietary, what I mean is, technology is only accessible to customers, when the customers pay for that technology somehow, with the open core itself, you know, quite literally, it is open, and it is licensed as such. And so anyone anywhere, would be free to seed, source, modify the source, contribute to it in any way distributed in any way. And there’s a few other dynamics to the sort of open the most commonly recognized properties of this open source software definition. And that’s the sort of general approach that we see the very large successful sort of hundreds of hundreds of millions in revenue, plus kind of commercial open source software companies implement as well as the early stage one. So you can think of the open core model as one where the company has the core open source project, or in some cases, a few or several, and the crust around or the show around that open core, where the proprietary software or technology is, is developed and distributed. And one sort of maybe over over simplified way of thinking about this, if you have sort of one open core and and one shell or one trust around it is the thickness or the thiness of that proprietary sort of shell or crust, varies depending on the relationship between the open source project and the degree to which the commercial open source software company can maintain or regulate a level of defense ability. And sort of the rate of commoditization, if you will, around that open core. And so this is something that has really just occurred and been true of commercial open source software companies for far more than 10 years now, which is that the most successful large scale revenue businesses in terms of product revenue in commercial open source software companies, occurs when the company is able to regulate and control the rate of commoditization around the open core. So around meeting the shell of the wrapper around the open floor. So to give you some concrete examples of this is a set of features, or extensions, or wrappers and there’s some, you know, common things that companies do, but it really just comes down to really anything that can be differentiating, and valuable for a customer. If those things are very easily sort of cannibalized or implemented or are built by the open core community and open core project itself, the open source ecosystem, or sort of anyone, the likelihood that the company or the commercial open source software vendor can maintain a differentiated sustainable, durable business, selling a commercial offering or commercial product to enterprise customers. It basically erodes and it diminishes over time. So that’s kind of one of the things that goes back to the relationship between the company and the open source project. And it doesn’t doesn’t necessarily need to be monopolistic, or sort of overly controlling on all dimensions, because each dimension is slightly different. But that’s that’s kind of one one way of thinking through the open core approach. And for us, this capital, our belief system, is that this open core model is, is one that works really well. And it’s certainly not binary, it’s not kind of fully understood. And it is it is fairly subtle and, you know, multi dimensional, and we think it’s context specific for each company. And it’s actually the main reason we decided to call this upcoming commercial open source software company conference or sort of industry event, the open core summit, because this is sort of the one common thing that the sort of commercial open source companies have in common. Is commercializing through through this open core approach?
Commercial open source software and venture capital
Yeah, I mean I think I’m always surprised that it took venture capital that long to sort of wrap their head around this open core model. I remember the Github 100 million series A led by a16z where everybody was thinking they’re absolutely mad. To be precise, Github is closed-source, but it’s mainly used for open source project develepment. And then of course, last year, we had this breakout year 2018 for commercial open source startups with GitHub, being acquired by Microsft, and in total 4 billion dollars exits. When I think about VC funds financing open source projects, some big name generalist funds and individuals come to mind, like Peter Fenton from Benchmark, Peter Levine or Martin Casado at a16z or Michael Volpi from Index. But at these generalist funds, commercial open source investments have always been part of a much broader portfolio. So, I mean, I’m really excited that we now have like a really specialized fund, investing only in this space. I remember the founder of Jfrog Ben Haim saying that, when they were raising their A, it took them like 4 years and then two weeks to raise the series D. So I think we’ve seen a lot of maturation in how VCs are viewing and understanding the open core model. But what do you think? Why did it take venture capital that long to sort of wrap their head around this new model?
I think that the the proof is always in the pudding. And when you see billion dollar outcome after billion dollar outcome, and particularly the last year’s sort of threat of M&A, IPO and private equity outcomes, across Magento and Elastic and Pivotal and Mulesoft and GitHub, and obviously the funding events like HashiCorps, hundred million dollar round, JFrogs hundred million dollar round, Gitlabs hundred million dollar round and so on, it just starts to become very obvious to kind of everybody up and down the institutional kind of capital supply chain, if you will. I think that Peter Fenton at Benchmark has always been super smart, you know, sort of active investor ever since making the leap of faith, investing in JBoss, I think 2003 2004 and I think he’s talked about this publicly couple of times. He’s definitely one of the more prolific what I would call commercial open source software company investors of the last 15 years. And Mike Volpi at Index has made some really phenomenal investments in this in this in this category. But as you mentioned, at the firm level, we’ve never seen an exclusive, concentrated focused thesis where that is literally all the firm does. I think for a few reasons, and I’ll say, one is just generally speaking, historically, venture firms have been very multi-thesis, sort of, you know, arbitrage, you know, multi-strategy, multi-thesis kind of multi-concept, because they want to sort of hedge against a new way or a new transition or trend coming along, and they being conflicted out or being restricted from investing in that. I think we’ve seen a big change over the last couple of few years. And that in that sort of world where there’s a lot of specialized firms, a lot of specialized funds, and I think abstractly, macro level, there’s going to be some really interesting some really interesting evolutions and the really specialized thesis focused firms over the next 10 years. And we’re certainly one. I think, there’s also going to be continued success from the multi thesis sort of thesis agnostic, if you will, venture firms, you could arguably say Andreessen Horowitz is thesis focused with sort of software eating the world. And going back to Marc Andreessen, seminal piece, but I like to think that open source software is eating software. So, you know, that thesis I think, evolves a bit, having said that, of course, Andreessen Horowitz has actually made lots of really smart, commercial open source software investments, and, you know, they’ve got great folks over there, in particular, like Martin Casado.
Commercial open source software eating software
Yeah, let’s dig a little bit into this seminal software is eating the world thesis of Marc Andreessen. And I think you’ve made the point before that, you sort of see in the CSS space, you sort of see that the open source community started to to sort of chip away at the middleware level. And then as a medium to long term perspective, you see it also moving into the application layer. Namely, SaaS, CRM, ERP, and maybe it very far out the consumer applications, basically having an open source Uber. Some, obviously the value there is in the data and the network. So that’s going to be a far leap. But maybe talk a little bit about this, this kind of narrative that you’ve mentioned in some of your blog posts already.
I think there’s a lot of nuance to open source eats everything, and open API’s, open interfaces, open data, open networks, a lot of these things are sort of overlap be intertwined, and somewhat related, but we’re particularly passionate about open source. And so I’m not going to make the broad panacea statement that hope source truly, you know, sort of eats and replaces every possible, you know, business model and, and company and technology and so on. You know, for the, for the sort of next several decades, I think that there’s there’s ways in which that could materialize from from different kinds of perspectives. But going back to what you were saying around sort of infrastructure, middleware, SaaS, consumer applications, and sort of that progression, all sort of, I’ll sort of give you one perspective that it’s sort of an analogy, way of thinking in terms of a very different sort of set of circumstances, and then sort of things that we saw in the in the consumer internet, world early on the Dotcom, bubble, and sort of 15 years later. And I think, transferably, what we what we could see with the developer ecosystem, open source and the sort of disruption at the application layer, beyond middleware, and IT and Dev Tools over the next decade. So this is this, this takes a little bit to unpack, and I tried to blog about this, once I failed, it was like one sentence, but it really requires like, 5000 words or something. So I’ll try and give you, thedescription of this. If you remember pets.com from the Dotcom bubble, sort of, you know, buying pet food on the internet, they raised it, they were one of the famous sort of like, I know, they raised several hundred million dollar deal. They raised several hundred million dollars, they, you know, crashed and burned and imploded and then they failed. And I think they failed for two primary reasons. The one reason that they failed, was that there was a very small relative to now, number of people on the internet, there’s a couple hundred million people on the internet, I think, perhaps less than that. The second reason was, it was cost prohibitive to ship, you know, the pet food across the country. So if you wanted to ship, so if you’re here in San Francisco, and I want to ship that food to you and you know, wherever you’re at that would be cost prohibitive, you know wanting to ship it to Boston or New York even that’s also probably constantly it, primarily because the logistics and supply chain and sort of shipping infrastructure for heavy packages was was not sophisticated enough to make that to make that efficient for the consumer. So those two reasons, fast forward 15 years or so. And we have Chewy.com, which is the exact same business model as Pets. com started started at the right time in history, essentially. But now we have, you know, several billion people on the internet. So the number of consumers changed. And then the shipping logistics supply chain infrastructure for heavy packages, costs dropped down dramatically, thanks to FedEx and others. Going back to the developer, open source world, I think to the application layer, one reason that is very, very sort of easily connected to the Pets.com story is for failure is SugarCRM. So SugarCRM was actually able to build a reasonably large business like close to 100 million revenue business or perhaps more. And they were originally a commercial open source based CRM application. And they actually we shifted to proprietary model. And I think they stayed that way. And they recently acquired by a private equity firm, and it was sort of plateaued out. But what I believe held SugarCRM back is is sort of two primary reasons that are inter-related, very similar to the Pets. com story. So when SugarCRM was founded, I believe it was about about 12–13 years ago, the number of programmers and developers building open source software was I sort of venture to say about an order of magnitude less than what it is today. So maybe 4 or 4 million. Right now we have 20 plus million, you know, 30 million, if you if you subscribe to GitHub’s numbers, which should be accurate. And I think the resulting sort of effect of that is significant, so it’s two dimensions: I think the proximity of the software engineers building application level, products is so far removed, given that the number of software engineers is, you know, in the sort of low single digit millions. And this is really, you know, people writing business logic for custom applications are glue, or, you know, trading finance things, they’re not building sort of enterprise packaged software category products that can be brought to market across, you know, lots, lots of customers and sort of marketed that way. I think the change from when SugarCRM was founded to now is obviously the number of software engineers building using and leveraging open source software as their primary tool for innovating and building and, and constructing things has grown by an order of magnitude. But I actually believe that number 20, to 30 million of what it is today is not high enough for sufficient empathy and proximity to the application layer, to actually disintermediate, the product founder. So with SaaS companies, which is the sort of independent software as a service category, where you see line of business enterprise applications brought to market. You typically have product, business oriented founders constructing the application architecture and model and workflow building and bringing together an engineering team to do that. And so I think that happens, because you don’t, you don’t have enough critical mass of software engineers to basically disintermediate that type of costly process, which requires high levels of empathy for the customer workflow, and extremely nuanced understanding of the domain problem. And so my theory, this is a far wild out hypothesis is that in the next decade or the next two decades, if we if we continue to see the compound annual, you know, growth increase of software engineers using open source and contributing to open source that we’ve seen in the last even five years, which is, you know, say between 15 and 20%, growing per year, and some years 50 plus percent per year five, zero, we will likely have by 2030, 2035, many hundreds of millions of software engineers using open source software and contribute and contributing to open source software as they build applications. And the consequence of that growth will be that the empathy levels for implementation of domain specific line of business enterprise applications just organically by the software engineering community without the need for a deeply domain specific specialized product person will be vastly higher than we’ve ever seen in history. And I think that will be a moment in time when we’re actually starting to already see that the cracks and edges and certain new application categories, that will be the moment where we’ll start to see open source innovation, and disruption, but a lot more innovation and the application layer. So far today, especially as is representative, this index that we’ve been updating for 5 or so years now, the vast majority of large companies that have been built at the commercial open source software company level, have been infrastructure, middleware, sort of developer tool, platform level companies. There are a handful of exceptions to that. So this company called Liferay, which is a very large and successful 800 plus person, you know, basically portal enterprise SharePoint essentially, alternative. And they also compete with Sitecore, to fully open source project. And they built a business and open core business around that they also have sort of like application categories that are sort of straddling both worlds, which are sort of middleware, infrastructure and application, things like CMS and E-commerce based frameworks like Magento, Acquia, Attomatic is the company behind WordPress, Acquia behind Drupal, Magento behind Magento, know, E-commerce, CMS. And so those are technically application categories, but they’re not quite full fledged CRM, ERP, sort of like large enterprise application categories that are dominated by companies like SAP, and Oracle and so on. So I think I think that’s kind of that’s my, that’s my sort of like, thinking around the phenomena of open source software, eating everything. But the fundamental sort of lever for that is that, you know, huge increase in the number of software software engineers building, building and relying on and contributing back to open source software.
Another really interesting example that you brought up recently was an example of Mattermost, which is an open source alternative to Slack. And this is especially interesting now, shortly after Slack’s direct listing. Slack obviously has a lot of proprietary data on communication that’s going on intrafirm and they’re not monetizing that to the level that they could. So I think, one reason why some companies and also open source projects are prefering Mattermost, is because the control of their data, maybe in anticipation of future monetization by Slack. So talk a little bit about this example, and how this fits into this transition that you described.
You know, I’m actually going to compliment my friends over at Andreessen Horowitz and Martin and his colleagues for writing a recent post saying that data data network effects are sort of overhyped and overvalued, and I think was actually a really great post. I didn’t quite agree with the first time I read, but I think over time, companies will differentiate on really valuable and frictionless, consumer experiences, by just iterating toward that and understanding consumer behavior, indirectly or directly without the need for, you know, exploiting vast amounts of highly sensitive and private information and data around consumer behavior and activity. And I think that Mattermost as a commercial open source software, competitor to Slack, fundamentally doing that they’re building a successful open source project that has, you know, won over the sort of minds and contributions of many, you know, hundreds of developers and the community and sort of compound benefits that they see from that are pretty, pretty significant. Can they can they grow into as large a company as slack? I think so. You know, if you have to sort of see a lot of the late stage effects play out for them. Slack is a, you know, likely to be a $20 billion company after after being public for a while, at least based on current market market pricing. And you know, Mattermost is series A funded company at this point. So it’s to be determined. But yeah, I think a lot of this is useful framing and, and it’s actually great that you brought up matter most because this is this is one of these early signs of an application level, commercial open source software company, and it’s sort of in a new application category. And again, a company actually growing and proving out that model works even now.
First public investments of OSS Capital
So let’s talk maybe a little bit about the concrete deals that you’ve been looking at, and the first public deals that you can announce. So I checked out Dev.to, which was I think your first public investment. It is basically an open source social network for the open source community. I joined their site today and made my first post today. So maybe walk us a little bit through the decision process, and maybe some, some other deals that you can already announce?
Sure, yeah so Dev.to is a deal that we were super fortunate and lucky to find and, and partner with the founders, in the first funding round, which was a seed round. You can think of Dev.to as a new kind of a content distribution, platform and social network for initially focused on software engineers and you know, sort of user generated content, that is very high empathy towards people who just want to go down a self-learning path and and, you know, really aren’t interested in taking lots of structured courses, and, you know, going through diversity and so on. Like, they’re really just self taught self motivated geeks or technology, people who want to learn how to write code and build applications themselves. And sort of the maybe the frustration, and friction in something like StackOverflow, where you’re just copying and pasting code code snippets, and expecting people to sort of fill in the fill in the blanks, with Dev.to and the Dev.to community. What you see is a much more warm, friendly, welcoming kind of environment that really connects with people who all sort of feel this shared sense of explaining the value of explaining explain ability, and the value of sort of demystifying concepts and breaking things down and accessible, understandable way. So I guess some concrete examples are like, you know, maybe when something like Webassembly or GraphQL, or, you know, Kubernetes or some new technology or new complicated platform or paradigm comes along, instead of sort of saying, ‘go read the docs’ or, you know, ‘go read the code’ or something, you can go to the Dev.to community. And, you know, invariably, you’ll find just like a litany of really valuable useful blog posts, actually, in understandable terms, breaking down what something does, how you can use it, practical kind of scenarios, and in a sort of simple and friendly way that I think has has has set the Dev.to community apart and made it a destination for some really awesome people. And it’s been growing really fast, I think, partly because the growth of the next generation of software engineers, I think, are primarily going to be self-taught, and primarily going to be going through these sort of online resources and tools that are just freely available and at their fingertips, but perhaps are missing the sort of human element of kind of explainability and cohesion, that the other platforms miss. It’s really interesting that you mentioned just sort of sharing content and publishing things as as an alternative to Medium or other proprietary blogging platforms. This is this is also another really interesting sort of growth factor for the Dev.to community, mostly because the platform is open source and the commercialization possible in lots of other ways then actually charging directly for content universally across the board in setting up sort of paid tiers and so on. And so yeah, the Dev.to, sort of content, publication, and the whole site, infrastructure, and CMS, everything is all open source up on GitHub. So you can actually build your own Dev.to for specific purposes. And dad is actively working with, with enterprise customers and helping them sort of build their own internal dev communities for their own internal software engineering organizations and helping them build platforms to connect internal engineering teams and so on. So it’s a really, really exciting company or crazy lucky to have to partner with them. Other companies we actually did invest in, in Mattermost as well. We participated in their in their previous round, and they’re they’re doing really well, we talked about Mattermost a little bit. We’ve invested in a number of other companies, over time we’ll be announcing a lot of those. I’m a big fan of talk about one who I’m a big fan of this is a little bit more technical. The RISC-V ecosystem is an open source, computer architecture, that is sort of set up to cause lots of innovation and disruption in the semiconductor industry. So the semiconductor industry is sort of done dominated by to computer architectures today x86 by Intel and ARM by ARM Holdings. And so Arm Holdings is microcontrollers, and embedded processors and sort of IoT and sort of mobile devices and so on. And x86 is all sort of server side infrastructure that primarily is only dominated by Intel, RISC-V is an open source computer architecture specification. And there’s a huge, you know, burgeoning ecosystem of companies growing around that. So we invested in in Hex five security, which is a security security platform built on top of this kind of computer architecture. And there’s a there’s a large, there’s a large ecosystem of companies building, building really interesting tools and products around that.
Investor value-add stack
Okay, so the last question before we have to wrap wrap up. Naval often talks about venture capital being advise being bundled with money. Not all money is the same green. So at OSS capital, I heard you talk about like sort of the value-add stack that you as a VC can have starting with licensing, where you obviously have deep domain access with Heather Meeker. Maybe talk to us a little bit about the value prop that you have when you’re when you’re pitching, highly competitive deals in this hot space.
Sure, how do we differentiate? How do we win over really great teams and founders? Multifaceted, but I’ll sort of stack rank things. Commercial open source software is all we do, we do nothing else. So we’re deep domain experts, we have a platform of equity advisory partners, many of the largest commercial open source software companies, and the founders of those companies are in that network. And so we can help with office hours and due diligence, and on our side deal flow and referrals, context-specific things with companies like product building, and hiring, strategy, and so on. We have a partnership, which consists of portfolio partners and sort of operational experts and domain domain experts on different dimensions. So things like finance and legal and product development and ecosystem development and so on. And those partners are more hands on and involved in our portfolio companies and also companies we work with, to help them build their businesses from early on. And those portfolio partners have helped grow and build many multibillion dollar commercial open source software companies from very early on, you know, pre-revenue all the way up to 100+ million in revenue. The another thing that we have as part of our platform is the open core summit. So open core seven, it’s kind of a vendor, vendor neutral industry event, we’re not really kind of tightly coupling OSS capital to to open course summit. The first event is happening in late September, we’re hoping to have well north of 2000 people at that event. And it’s kind of the who-is-who of the largest founders, commercial open source software companies, as well as a large contingent from the investor ecosystem analysts, aspiring founders, developers, and so on, as well as customers and sort of you can think of that as the commercial open source software industry ecosystem conference. And that’s really kind of a place to learn, understand best practices. And really part of the semantics of this category is that we review and we believe commercial open source software companies are fundamentally different in almost every functional domain. So product engineering, hiring, business models, finance, and so on, as compared to proprietary software companies are fully fundamentally proprietary software companies sufficient to really benefit from the deep domain expertise and exclusively focused kind of knowledge in this area. And we’re super happy and , you know, ecstatic to partner with great firms, like Andreessen and Index and Redpoint and others. We’re a very early stage focused firm, and we don’t have you know, 10 billion under under management. So we’re just kind of, we’re less than a year in and we’re getting started. But hopefully, over time, we can we can add a lot of value and help help companies that are excited about this model.
Where can you find out more about Joseph and OSS Capital
Yep. Thank you so much, really excited to have you in this in this ecosystem of really specialized domain experts, investing in this space. And so maybe how can people get in touch with you and find out more about OSS capital?
Sure, you can go to our website, it’s it’s really simple. It’s just OSS.capital. And for the open core summit that’s coming up the end of September, September 19. To 20th. In San Francisco. It’s just opencoresummit.com. I’m pretty active on Twitter, I’m @asynchio on Twitter. Just you know, put in Joseph Jacks on Twitter. Otherwise, yeah. Thanks for having me on Erasmus.
Thanks so much. What a great episode this was. I don’t know about you. But I personally learned a ton about source software companies, and how this open core model is really challenging this traditional, fundamentally proprietary software model. So this is it for today. Make sure to tune in next time. And make sure to subscribe to my newsletter on sandhillroad.io Cheers, guys.