The Business of Open-Source: A Conversation with Heather Meeker (Part One)

Forbes Books presents the Startup Science Podcast with Gregory Shepard, brought to you by startupscience.io. Greg is an entrepreneur who has built and sold 12 businesses. He's a recipient of 4 private equity awards and is a featured TEDx speaker and Forbes author. Here's Gregory Shepard.

My guest this week is Heather Meeker, a leading expert on open source software licensing and business strategy. She's the general partner at OSS Capital, an early-stage VC fund exclusively investing in commercial open-source software startups. We're also going to talk about her latest book, From Project to Profit: How to Build a Business Around Your Open-Source Project. Heather, welcome to the podcast.

Oh, thank you. It's a pleasure to be here to talk to you today.

When we look at open source and Building and scaling open source software businesses can seem like magic, right? And at least smoke and mirrors in some cases where you're, you're talking to open source people and they're trying to get people to understand what open source is in the first place and how that could even become a business. So I think to start out with, just to get everybody leveled up a little bit here, over the last 30 years in open source, you're a massive specialist in this area. One of the questions that I get from people all the time is how is open source a business? Like, how does— how do you take something that's where engineers are developing code that's available for anybody to use and turn that into a business in the first place?

Yeah, there are 2 things to keep in mind about that. It can kind of seem like magic or smoke and mirrors, but if you take a step back and you think about a couple of things, it makes a lot more sense. One is that If we look at the way business works today, as opposed to maybe the way it worked 10, 20 years ago, we've got a lot more businesses that offer a lot of free stuff and then some paid stuff. We're all used to that now. It's mainly because of the advent of the web and, and the advent of digital goods and, and technology. It makes it easy to make stuff available to people to try out. And so we're all accustomed to this kind of free goods model, and that's part of what's going on with open source. The other thing that's super important to understand, and I actually think even most people in Silicon Valley in the tech business don't really get this, so I love to talk about it, you know, is That open source, it's a license to do things with source code, an unrestricted license to do things with source code. And those are intellectual property rights. It's basically a grant of a copyright license. But nobody sells intellectual property grants. It might seem like people do on the surface, but even if I— let's say I buy a subscription to a software product. I'm not really buying the intellectual property. What I'm buying is a product that does something for me. And there's actually a delta between, say, the source code for that thing and the packaged-up, served-up product that does something for me. And the difference between those two is mostly what open-source businesses sell. And it's actually, it adds an enormous amount of value, between just kind of the bare intellectual property license and the actual product that's put together that people can actually use to solve problems.

Mm-hmm. And you would know as somebody that was named Business Insider and one of the top 10 people transforming the way technology does business, along with the CEO of Salesforce, Stripe, and Microsoft. I mean, like, that was like my little surprise for the listeners. I'm sure many people are like sitting up in their chair now going, oh, wait a minute, wait a minute. So when you look at this open source community and you look at what's happening with AI, do you see AI starting to move into more of an open source environment, or do you think it's going to be more privatized?

So I personally think that unless AI becomes quote unquote open source to some degree, it's going to have a lot of problems. The thing about AI is that we really don't have models for what open source AI means because AI, you know, machine learning models are not source code. They're a different creature. And we have this great model for open source and open source licensing that we've been using for like 30 years and everybody loves it and it's proven its worth and everything. But it's very hard to graft that onto AI because of the difference in the technology. So just to kind of explain what I mean, if you want to give someone the ability to study a software program and see how it works and maybe think about improvements, you give them the source code. And if they're skilled in programming and so forth, they'll get there. Like, they'll look at the source code, they'll be able to Figure it out, see what changes they want to make. Machine learning is different. It's not even clear what you really— all the stuff you really need to give people in order to make models transparent and improvable and so forth. And the industry has not figured that out yet. So we're really in a state of flux with AI where a lot of people want to apply the open source model to it. But it's actually a little bit difficult to figure out how to do that. I've been spending a lot of time talking about that and so forth for the last year, ever since this stuff just blew up about, you know, a year ago. And people are working on figuring out what it means. Personally, I think that unless we can figure out a way to make machine learning models more transparent and accountable, Um, it— they risk getting basically regulated off the face of the earth or being just in the hands of really big companies that can, I don't know, like, uh, figure out how to make people accept them notwithstanding their lack of transparency.

Or to your point, the transparency, for the listeners to understand, is that machine learning goes through processes where it starts thinking and it does things.

Right.

And then it comes out with an outcome. And the part that everybody seems to be worried about is how did it get to the outcome?

Exactly. And it's, it's not, it's not a set of instructions. I think that's something that, like, most people just are not focused on. Like, with software, there's a set of instructions that was written, you know, by a person, or at least checked by a person. And so if you want to see how did it get from A to B, you can actually go through and follow the logic. But that's not how machine learning works. Like the transformational models, they just kind of, you put something in, it's kind of a black box and it comes out the other end. And that's of great concern to people because when you try to figure out, well, did it make this decision properly or fairly? It's actually quite challenging to figure out. Yeah.

I mean, it's incredibly challenging. And One of the fears that I keep hearing about, people, people keep saying is, you know, is this going to make us dumber? It's basically doing the thinking for us, right?

Like, yeah.

You know, what you would normally delegate to somebody, you're like, hey, can you go figure this out? Can you go do this research and come back to me with this? That's the kind of work that AI seems to be used for right now quite often, right? And I mean, we use it, right? But then, you know, you can't help but to think, well, is this going to make us dumber because nobody's actually learning? Because the AI is doing the work for us. What are your thoughts on that?

Yeah, I don't— I'm on the fence about that. I mean, there are a lot of things that AI can take some burden off of us for doing repetitive tasks that humans have had to do. One of the examples I use because I'm a lawyer is that like maybe a little bit before I started practicing, so it'd be in, you know, the late 1980s, When people negotiated contracts, they would— you would send someone a document showing the differences between one version of the contract and the other. It's called a redline in legal practice. And people used to do those by hand. And today no one would ever do one of those by hand. There's like programs that do them. The programs are more reliable and probably more trustworthy than the humans that were doing them. And so there are a lot of tasks like that that I think it's actually great if we can free people up to do tasks that require like more sophisticated mental, mental energy. But on the other hand, the machine learning models are very good at spitting out what I would call kind of plain vanilla prose. And one thing I do worry about is that people won't learn to write anymore. And, uh, and that's been kind of a problem, honestly, for those of us who have been in sort of management, uh, positions in— for knowledge workers, right? Is that, um, it's hard to find people who write well, you know? And I think it gets harder and harder every year. And I'm actually kind of worried that people, they're just not going to learn anymore because they don't have to What the way you learn to write is just by producing a whole lot of writing, right? You have to do it over and over again. So yeah, it'll have good and bad effects, I'm sure.

So let's talk a little bit about this book. So you have— there's a lot of different themes in the book: software entrepreneurship, investing, demystifying the process of building a scalable open-source business. And I wanna talk a little bit about that. So I wanna shift topics a little bit into that. One of the things that I learned is when I was looking at the content surrounding open source, and that I hear all the time, is why should somebody start an open source business? Why is this a good investment? And what does it come down to? And maybe you could shed some light on some of that.

Yeah, I'm a VC as well as a lawyer, and I specifically invest in commercial open source. Software development. And I think it's a great investment for many reasons. I'll just, you know, touch on a couple. The first is that when you're an investor and you're looking at, say, a company you want to invest in, one of the things you do is you go through a diligence process to try to figure out, well, how good is their technology and how, you know, how good is the market in Are people gonna be interested in the product and so forth? With open source, you can tell a lot about that stuff before you even engage with the company, because usually what's happened is they've thrown an open source project out there and you can see, oh, it has X number of stars, it has X number of contributors. I can actually look at the software and see if it's any good, you know? So it actually de-risks the investment a lot if you're interested in investing in a business that surrounds the open-source software. On the entrepreneur side, open— and the investor side, like, open-source businesses are incredibly capital-efficient. They, you know, long before the pandemic, open-source collaborators were collaborating from across the globe. They were virtual. They, they don't need offices. They, they run lean and mean. And that's the kind of business that's good for both investors and entrepreneurs. So I can give a business half a million, $1 million, and they can do amazing stuff with it. And I can figure out in advance, like, whether people seem interested in the product. And that's awesome, I think, for both sides. Anytime you have something that allows somebody to start a business with less capital, it means more innovation. And the capital that the investors provide gets used, you know, with more leverage and more efficiency.

So essentially, like, what you're saying is, like, you can earn and learn about product-market fit.

Yeah, exactly.

You can learn about the personas that you're dealing with and the ICPs that are interested and all the little things that present risk if they're unknown, because those things have already been launched and the numbers are already in. Is that—

Yeah, exactly, exactly. Um, and you know, I, I hear people say all the time, well, GitHub stars are kind of, you know, shouldn't pay that much attention to it. And I understand that, but, um, if you look at how many contributors there are to a project, how many downloads there are of the project, how engaged the community is, you have amazing intelligence about product-market fit. And that kind of thing is hard to buy with a regular business. So it's really exciting.

Yeah, I mean, and anything, you know, the naysayers that are out there, they always, you know, there's always people out there that are saying, well, you don't get this and you don't get that. And you're like, well, It's better than nothing. How about we start with that?

Yeah, exactly. Another great thing is that open source businesses get a lot of really good feedback from their communities. So, you know how you use products and they're always asking you surveys and all this horrible stuff, like how would you recommend this and so forth? Well, the open source projects get tons of feedback from their community. And that stuff is incredibly valuable, and you get it for free. Like, you don't even have to send marketers after people to talk about it. So there's just a lot of aspects of running a business around open source, if you, if you do it properly, that, you know, it's, it's a, it's a huge benefit to the people running the company and therefore their investors. Yeah.

I mean, you get the engagement right out of the bat. Yeah, because they're participating, right? So they've got skin in the game and they're willing to engage. And you're right, it, it, the, you know, how many times a day do you get the NPS email, you know, the Net Promoter Score email? It's like, would you recommend? And you're like, oh my God, you know, and you get like 30 of them a day and these different surveys all the time. And, uh, I think people abuse the hell out of that process.

But yeah. And in open source, people will open an issue And they'll do it for you, right?

Yeah, they'll open an issue and somebody else will solve the problem.

Yeah.

It's like incredible.

Yeah, sometimes.

You're right, it's a pretty innocuous process.

Yeah, the corollary of that though is like for open source businesses, they have to pay attention to their community. Like you can't just start a business around an open source project and then expect your community to just give, give, give to you. You have to give back to them. You've gotta respect them. You've gotta pay attention to them. You gotta take their concerns seriously. But if you do that, it's a wonderful partnership between the community and the company. And you can, you know, use that to leverage what you can do with the commercial product end of it. And you can use the commercial product revenue to leverage what you do with open source. And that's kind of the dance that you have to do with a business like that.

So, let's talk, we have a lot of hackers that listen to the show, right? From the different Hacker Angel Hack and all the other communities. And so, what should developers consider if they're thinking about going into the open source business? What should, some of the things that they should be thinking about?

First thing they should think about is whether it really should be a business. So, this is one of the things I talk about in my book. And the answer to that, there's no wrong answer to that. It's just that based on the answer to that, you go in different directions. So for instance, if I'm a developer and I'm scratching an itch, you know, I'm thinking this thing really needs to be written and I'm going to sit down and write it and spend my, you know, hours, nights, and weekends doing it. And I do that. I might do that because I just wanted to do it. Maybe I did it because I wanted to kind of put it on my resume. I might— it might be a hobby for me, and then it might start becoming a business. But even if it does, it might be in different levels. Like, do I want to start a business where maybe I'm going to become a professional services provider, do some maintenance and support for people and so forth? Create a living for maybe me, maybe a couple of my friends, and that's fine, right? But then there are some people who really want to take the open source project and turn it into a big business. So depending on what you want out of it, what your time horizon is for an exit, how big you want the project to be, do you want to actually manage a business or just want to work for yourself? Based on that, you get to the answers about how are you going to run the business? How are you going to fund it? What are your goals? So that's, that's one important thing. The other is that you have to have a business model for what you're going to do. And it doesn't need to be perfect. You just have to have an idea. If you go and you ask investors for money, you say, I have this project with 10,000 stars and it's really great and people love it, the next thing they're going to ask is, okay, what are you selling? Who are you selling to? How are you going to make money? And you need to have answers to that, even if, even if they're just your best guesses. But you have to have answers to that. And I think that some open source developers can tend to engage in a little bit of magical thinking about, well, stars means money. It's not— doesn't quite work that way. But you can get from A to B if you plan it out and you think about it.

Mm-hmm. Yeah, I think that's where I was coming from with the comments of, you know, black magic and smoke and mirrors. Because a lot of the engineers, you know, when I look at those deals, you know, they come in and the engineer's like, well, I've got 5 stars and, you know, I've got all these people contributing. And you're like, yeah, they're not buying anything.

Yeah, right. That can be a fine outcome, but it's not the outcome that will get you funding, right?

Right.

So you just have to think about like, what is your objective? What do you really want to get out of the project?

So I wanted to switch things up a little bit here. We talked about founder engineers that are creating all this stuff on one side of the equation. Now I want to shift to the investor side, but we have to stop here. Also coming up next time on the Startup Science Podcast, Heather will show a good vertical to dive into when it comes to open source.

If you had to pick one business model that tends to work really well with open source, it's, it's what we call an open core model backed up by SaaS. So what you do is you develop some core functionality and you release that as open source, and then Around that, you build software as a service.

The Startup Science Podcast with Gregory Shepard is brought to you by startupscience.io. Founders can log on to startupscience.io and use code FORBES to get free access. To connect and find out more about Greg, go to gregoryshepard.com. The Startup Science Podcast is a production Forbes Books.