Tyler Jewell is the CEO of WSO2, maintainers of a popular open source API Management Platform. In this episode, Tyler discusses WSO2’s subscription-based business model and the future of cloud versus on-prem services.
Michael Schwartz: Welcome back to Open Source Underdogs, the podcast where we get you into the middle of open source business models.
Our guest today is Tyler Jewell, CEO of WSO2, and a partner in Toba Capital, a Bay Area VC fund.
Tyler does a pretty good job of explaining what WSO2 is about, so let’s just dive right into it.
Tyler, welcome to the podcast.
Tyler Jewell: Hey, thank you for having me.
Michael Schwartz: I’d love to hear a little bit more about your background. You worked for Oracle, Quest, Veritas, PA – how did you get here? And what did you find so exciting about the opportunity to lead WSO2?
Tyler Jewell: Well, I’ve started off my career with a degree in Computer Science, I had a very short-lived career as an engineer. I wasn’t all that good with that, but I got this amazing exposure to enterprise software, spending six years of BEA in various product-related roles.
And that really led to a career in product management. Got exposure to different companies and different products, usually related to developers, middleware integration, DevOps, and successively worked on projects of different complexity and scale. And that really helped build a global view of markets, products, and how you move those products to those markets.
And, through accident more than anything, while I was at Quest Software, and doing product management, they had a view of markets, which were combination of mergers and acquisitions, investment, and product management in the classic sense. And I got invited to work on a variety of corporate development activities, a few acquisitions, some divestitures, and for a while, leading their investment portfolio.
Once you start wandering in the corporate development, you really develop a broader view of markets, and you start to look at things about, “how do I help develop the market?” And there’s lots of ways to do that besides just building products.
And that just gives you a completely different perspective on the ecosystem, how businesses are built, and that opened up all kinds of other opportunities for me. I started doing board work, I had done angel investments, the corporate investments.
Over the years, I put about a hundred million dollars of venture capital and money to work in various companies, all related to DevOps. That got me opportunities in Oracle, and eventually allowed me to form the hypothesis on why I started coding.
And it was never a goal for me to become a CEO, but just the opportunity was so exciting, and people were asking me to lead it and go forward with it. Next thing you know, you’re the CEO of a company. Codenvy, the investments were just a wonderful experience over 6 years.
And when Codenvy was sold to Red Hat last year, the founders of WSO2 – which I’ve been on the board for almost eight years – were ready to focus on being technical founders, like, the relationship that we had, and asked me to come on board here.
So, it’s just been really one thing after the other, there’s been no particular vision in mind other than good technology, good people, and kind of developing markets.
How Does Open Source Enhance The Business Model?
Michael Schwartz: WSO2 has created a number of successful enterprise integration software platforms. How do you think that making these platforms open source enhances the business?
Tyler Jewell: I feel like that open source is this important trade-off that you make. By opening your source code, you are creating an environment of visibility and transparency that should lead to a community in all sorts of ways that you can’t even anticipate in the beginning.
So, if you can put value on some definition of the community through that transparency, then it’s a good trade-off to make. When you’re dealing with investors, I tell investors, “look, open source really is supposed to bring investors to key attributes.”
The first thing is that if the investors are paying for the intellectual property to be developed, and then the company gives up some of these rights by putting an open source license on it, then in exchange for that, they should get a significantly higher rate of distribution on their software products.
If it’s available, and there’s limited restrictions on access, they better be paired with some sort of mechanism, by which you get a hundred, or a thousand, or ten thousand extra distribution. And that’s meaningful because it should generate a lot more opportunities for you to sell your value than it would have otherwise.
The second thing is that there is this unwritten, implied contract between the open source vendor and its community. And the community, whether they realize it or not, because they have access and rights to the software, they do take on some implied obligation of evaluating the software a little bit more thoroughly, without depending upon the vendors resources.
So, there should be, overall, a much lower cost of sale because more of this evaluation process can be done without the vendor having to commit resources to do it. So, from an investment point of view, those two qualities really should materialize, and obviously that should leave to a good rate of return.
Michael Schwartz: WSO2 has several products, and I guess each product has its own specific value proposition. Do you think that there is an uber value proposition that is common among WSO2 products?
Tyler Jewell: The perception is, as WSO2 does have a number of different products, but we play in a single market, which is the integration market. And the integration market is really defined itself to be API-driven over the past few years.
So, you don’t do integration unless you’re going to package it up into some sort of API. And so, every product that the company has is really tailored around this narrative on what an integration platform needs to take in order to deliver on this API-driven vision.
You start with APIs, you then need API management. In order to build the APIs, you need to have an integration backbone, so that’s an ESB and message broker. And then, you need to secure all that, and that’s how we get to the identity space.
The company, for a long time, was very technically-oriented, it positioned a lot of these things as separate, individual components, but all the components were really designed around this common integration narrative. Over time, I think the company has gotten better about how to communicate that.
What Does Open Source Mean To WSO2
Michael Schwartz: What does open source mean to you? Is it just licensing the code under an open source license, or does it mean something more?
Tyler Jewell: Open source to me is a commitment to openness more than anything.
I think that the phrase “open source,” it too easily implies a licensing model associated with that. But if you’re really going to call yourself an open source business, you need to have a commitment to openness. And when you commit to openness, this creates a level of transparency that allows you to get a level of trust with your potential customers, and when you have a higher degree of trust, your interest becomes more aligned along the way.
I think that most companies start off by just looking at this as a licensing mechanism, they can put it out there, and you get this kind of distribution exchange. If you start to see that the whole business process and the workflow around the intellectual property can be equally opened up, it just really changes the dynamic of how your prospects, and your partners, and your employees engage with each other, and with the rest of the world.
That creates these aligned interests, which, then, fuel growth in profits, which are the things that businesses look for.
What Types Of Software Should Be Open?
Michael Schwartz: Not all software necessarily should be open source. What types of software do you think lend themselves to being open source?
Tyler Jewell: I’ve certainly seen it, the case where it’s an enterprise software, where if, there is a wide spectrum of opinions on the direction that the software can go, opening that up to community is a very effective way to try to rationalize and deal with all those different perspectives.
If there are lots of concerns about the configurability, or the security of the software, having it be open and transparent gives almost kind of outside credibility, and mitigates the risks that could come from very complicated software.
So, we’ve seen time and time again that that tends to work very well, when you have a consolidation of the commons of reconciliation of wide-viewpoints – open source attempts to do very well there.
I’m not all that well-informed about the consumer market or the pre-packaged software market, but it strikes me that we’ve seen less success, at least commercially, with technologies that take highly shrink-wrapped products, and then open source them, where consumers just expect them to work. It either works or it doesn’t, it’s a very binary sort of thing, and there’s not a lot of engagement.
I think we’ve seen from overtime that when products fall into that very tight sort of packaging, the community doesn’t give as much respect for the nature of the open source, and I don’t think there’s a fair equilibrium that shows up on that, so that my sense is that they’ve done less success in that regard.
Michael Schwartz: You announced that WSO2 will be launching more community-driven, open source initiatives in 2019 – how are these different from WSO2-driven initiatives?
Tyler Jewell: Yeah. What’s interesting is, WSO2 is 14 years old now, and when the company started a long time ago, their vision was certainly working in the integration demand, but they were building out kind of this big platform, singular platform that was an alternative to websphere and weblogic at the time.
It was really the WSO2 product, and obviously, they called it a lot of things, but the platform essentially was the product, and it was this big thing how to solve big problems.
Since then, all the products that we ship are based upon this Core Kernel that we’ve got, and it’s all WSO2 branded products that’s there. And there’s no concept of an upstream with us.
The WSO2 repository builds the WSO2 product, there’s no forking, there’s no branching, it’s a very tight linear relationship that’s there.
People who tend to adopt our open source, first have to become familiar with our brand, build some reference ability, and see that we’re a quality brand, and then they’re willing to adopt the WSO2 product line in whatever configuration they think is fit there. So, it’s a kind of a very direct, linear relationship.
What we’ve seen as very successful over the past seven years is that the nature of marketing and how products are evaluated have shifted, and that you can take any number of critical and essential components to enterprise architecture, put them out as open source technologies and build community and transparency around those, and under a different brand.
It doesn’t have to be your corporate brand, it can be under some project brand, or whatever it may be, but the important impetus there is used by your peer group.
Can you get other developers and technologists to adopt it, can you collaborate together on those things, and through that, you gain adoption and interest. For us, we’re going to be committing to more of those types of projects, and then using those components along with components from other third-party, open source providers, rolled into our products, if you will, the WSO2-branded products that we can provide support and commercial capabilities around.
I think that good companies today have a blend of these products and projects. And WSO2 has predominantly been almost only products, we’ve had some projects along the way, like, Ballerina and City, but we’re moving towards having bigger commitment to these projects along the way, whether there are projects or third-party ones.
Michael Schwartz: You don’t charge for the right to use this software, and I guess you probably get this question a lot… So how does WSO2 make money?
Tyler Jewell: We don’t charge for the right to use the software, we sell subscriptions. Subscriptions are the mechanism by how we deliver most of our value to our customers.
And subscriptions come with a couple of things. One, it gives you the right to ask us questions in development or production, it gives you an SLA , and when you’re dealing with very complicated software that’s powering trillions of transactions, now you need a response SLA on this particular issue.
It also includes updates. The updates are really what make us unique.
I think that we have an approach here that is unlike any other open source software company.
With the updates, the updates are where you expect patches, fixes, security issues, things that were unexpected when we had released software. But what makes this particularly unique is that when we need to do a patch or a bug fix, we still release that as open source, it is still Apache licensed.
And we put that into master, and it’s in the unreleased version of the repository, and it’s available for anyone to get.
So, someone can compile that, and they can put that into their environment. But the service that we provide to our customers, who are on a subscription, is, we take that patch, we backport it specifically to your version, we convert it into a binary so that you can install it, just that binary and nothing else. Then, we also test that binary pretty exhaustively, so that we have the highest confidence level that what we’re giving to you is going into that environment.
What we do for the customer is, we take the patch, we backport it to the specific version for that customer, we convert it into a binary, we test that binary on environments that largely mimic that customers environment, and so we have every level of confidence and risk mitigation that the patch that we give to the customer is going to be meeting the needs and expectations they got, without disruption to them.
And those binaries that we ship, those binaries are only available for those people on subscription.
And that’s how a big way of how we communicate that value, we get to hold to our premise, our open source principles because everything is still technically open source in the code basis, but the value that we provide is in the work that we do with the binaries and patches and the risk mitigation of that. And our customers look at that and say, “Oh, you know, well, that seems like an equitable exchange of value here. We’re happy to pay the subscription.”
Michael Schwartz: I was reading in your annual report that you have an excellent track record with customer retention, so that subscription and providing the binary updates provides an incentive for customers to renew. Any challenges around getting renewals?
Tyler Jewell: If you look at our dollar-based retention rate, it’s equal to Mulesoft, which is a mostly proprietary platform on that, so we’re doing well in our particular peer group.
When customers turn on us – we do have some customers who just downgrade to a peer open source, for whatever reason, maybe they had budget cuts or whatnot – but we look at those not as negatives.
I mean, I think that on a dollar basis, you can look at it and look at it as a kind of a loss, but when we go into those counts, we say, “look, you know, if you’re going to do this, you still want to be active in the community, you can still act as a reference to us, you’re still happy with the technology, you can still give us feedback.”
So, they’re still strong contributors to the community overall, and there’s an immense value that comes with that, even if they’re choosing not to get the updates in the SLA associated with it.
And when we look at all of our turn, and every company has it, about two-thirds of our churn are companies who are sticking with us, and so it’s hard for us to look at those as losses.
Michael Schwartz: So, one question I have is about – and I don’t want to get too deep into licensing because it’s a really deep topic – but sometimes customers, I get the comment that customers know how to purchase a commercial license because it comes with reps and warranties and acceptance of liability.
But the open source licenses normally try to minimize liability and minimize reps and warranties – does a subscription address some of those issues too?
Tyler Jewell: Yeah, we, with our subscriptions, can offer different types of identities and warranties for customers. We do it both, for the patches in the binary form that we issue and for the open source code.
We have a little bit of an advantage there, we hold copyright to everything that we write ourselves on that. Even in the situation where we get outside contributions, we take copyright. And when we hold clear fork so that we can maintain that, and do all of our legal checked.
That sounds like a control provision, and it’s not so much of control provision, it’s more of you want to have complete and total visibility to the entire code stack so that you can patch any portion of it all the way down, even if necessary you need to get into the JVM, or to the OS as well.
Once you can demonstrate that, on the legal basis and the business basis, you can start getting more flexible with the legal terms that you negotiate, and on top of that, you can make higher commitments on the SLA basis. And we’ve even offered truly abnormal response time SLAs, for customers who feel that they need that.
Michael Schwartz: One of the WSO2’s priorities that you documented was to build a culture of transparency. I have to say you’re living up to that. WSO2 is really transparent. I think we’re private business – why do you think transparency is important?
Tyler Jewell: Transparency is really essential to our business, we practice it in number ways. I think it’s fundamental.
I think transparency is the most effective way to get diverse groups of people with different points of view to be aligned. And it does that because when you’re transparent, there is independently verifiable information.
And when outsiders start to absorb that information and can verify that, it allows them to build trust much more quickly. And when two different parties have trust with one another, they are able to align their interests much more quickly, and then collaborate.
This works well, not just with meritocracy, with the Apache way in software products, but we found that it works well with our internal culture – with how we prepare our marketing and sales,
how we treat our financials, we are actually pretty transparent about our financials outside and online, our hiring practices, our partnering practices.
I’ve just found that as the larger that you get, growth really comes from the momentum that you can establish. And when you’re dealing with large populations of people, and product, and partners, and customers, you can build a lot more momentum by getting everybody’s interest aligned and transparency is the fastest way to do that.
Michael Schwartz: WSO2 is hiring a lot of people. The goal said 175 people in 2019. That’s almost one person per business day. What are some of the challenges that come along with this velocity of growth?
Tyler Jewell: Well, the biggest challenge is, you need to have a source of talent that you can work with. And we found that there are really two great ways that you can find the source of talent.
First our largest office is located in Asia, and it’s in a region that has very high-quality graduates in computer science, and computer engineering and business. It’s very important for us to be able to be credible and well-respected, so that we can work within that community and do a big portion of our talent through that.
The second is that you have to really look to everybody, all your employees.
Your employees become your recruiters because it’s an environment that has a certain set of ideals and principles, and they’re working out in other communities that have people who reflect similar sorts of ideals, and so it becomes really essential that everybody is recruiting other people that they really respect along way. The source of talent is one really important thing.
The second that’s really challenging is culture and onboarding. In an open source company, so many of the people are mission-oriented. They work for the company because they have a high degree of aspiration to the outcome. They work on it for the love of the technology, they work on it for the brotherhood with the people that they get to work with. There’s a lot of that that goes on, and less so on a compensation or a financial incentive basis. I think we all work for money, but this mission orientation, this mission theme is a big part of that.
When you hire larger and larger groups of people, it becomes more challenging to get everybody aligned on that mission, and for everybody that you have an equal share of the commitment to the principles, and so you really have to start investing almost exponential higher rates, in terms of culture, as you get bigger.
Otherwise, you’re really going to start to fray at the edges, and the talent that you get isn’t going to stick around, you can have a high degree of attrition, and long-term people are going to start to lose sense of the mission as well. So, these are the two big things we face.
How To Build Culture
Michael Schwartz: When do you say invest in the culture – how do you invest in a culture exactly?
Tyler Jewell: First, you need a really clear and concise definition of your principles and ideals, and it can’t be marketing fluff. It’s got to source itself from the founder’s beliefs and the founder’s behavior.
Then, you need to have a commitment that everybody’s going to live up to those principles and ideals. You have to be willing to tell people that it’s okay if they don’t believe in this particular approach, or it’s not comfortable for them, that that’s okay, that this is maybe not the environment for them.
So, how do you go about doing that? Well, you got to have a lot of materials to explain it, there’s a lot of collaboration that you need to do with your tech leaders and your middle management across the company, so that they can embody this and practice it. We actually incorporate our principles and philosophies into our performance review system, and that’s all part of everything that we do day-to-day.
And since we’re so community-driven, and it needs to be focused on the mission and the theme, you have to invest a significant amount in people and things that allow people to participate in that community, whether it’s games, company all hands sessions, big office spaces that allow people the flexibility of their work schedule, whatever that is, everybody participates in the community in their own way, and you have to not make that equippable budget item.
That has to be kind of first on the budget. You have to put all those things in place, so that those people, who want to be a big part of our community, whether it’s our internal/external community, have every tool and resources available to them to do that, and they can’t point anything that’s holding them back on that.
That’s a big part of it as well, as steering increasingly larger numbers of resources, and time, and focus, into those activities.
Michael Schwartz: Switching gears just a little bit, WSO2 recently announced some new cloud offering. I’m wondering if you think cloud will come to contribute more revenue than software subscription in the future.
Tyler Jewell: You know, I think that Cloud is a delivery mechanism for sure. Is it going to overtake on prem versus not on prem, I’m not sure that I’ve got a particular bias on that.
What I do see happening is that Cloud is definitely changing the velocity at which new services can be brought online. I see the services cropping up on prem in the cloud hybrid everywhere, and so I think we’re looking at a truly hybrid world.
And that services are going to be able to reside anywhere, they’re going to be able to live anywhere. From our point of view, we think that that creates a wonderful integration opportunity, a lot more things that need to be brought together into a common view.
Which one of those things wins it’s hard for us to say because each and every environment has such unique and different needs. But it’s hard not to ignore the pure scale and size of Azure, and AWS, and Google – pretty impressive revenues across the board.
Experience From Cloud
Michael Schwartz: Has moving to cloud services created a friction in the type of business that you were? Because it seems, like, in the cloud business it’s about operations, whereas in the software vendor business, it’s more about customer relationships and product innovation. Has there been any growing pains of moving into this cloud business?
Tyler Jewell: We definitely had a lot of growing pains there.
Right now, our public cloud is really just a public cloud variation of our on prem products. So, it’s not much of a hybrid situation, the value propositions are largely the same in that regard.
I think that what we’ve really figured out is that in order to run a public cloud, we had to operate our products in ways that we never operated them before. And that, in turn, caused this to redesign a lot of basic primitives in our products, related to update, configuration management, operations, and whatnot.
And that, in turn, is actually one of the big reasons why we’ve had the kind of growth over the past couple years, because 35% of our customers are running our products. Even though it’s an on prem version of our product, they’re running it in a cloud style.
So, we basically ate our own dog food, and going through that process actually caused us to rethink all sorts of things at the end of the day. That was probably the biggest benefit we got.
Michael Schwartz: What’s the typical sale cycle? Do people find the open source, and then call you? Or is WSO2 more active in generating business?
Tyler Jewell: We are more of an inbound business, believe it or not, because we put our technology out there, we evangelize it pretty aggressively. So, customers tend to approach us, and they’ve already got a project in mind for that.
They approach us in one or two ways. About a third of our customers approach us through a partner, so, a system integrator or reseller that they’re already working with, and a lot of those partners are equally committed to open source, and we are their open source vendor of choice in that regard.
Then, other customers have architects and VPs of engineering, who have been designing systems, and they see that we could fill portion of their architecture, and they contact us.
It’s almost always the case when, by the time they contact us, there’s done a lot more evaluation than you might be surprised.
And they’re contacting us because they’re looking for advanced system design, to hear the total story maybe they’re ready to engage on a subscription. It can go any number of different ways, and that process takes about 3 to 9 months. With most customers, sometimes it’s up to 2 years, sometimes it shorter than that.
Customers come with us, they come to kind of really two starting points.
One starting point, which is obvious, we have a lot of different components, is they’ve already got an architecture, the architect is very confident that they’ve got the right design, and they’re just shopping for certain components. “I just need an API Gateway.”, or, “oh, you know, I really need a message broker.” and they’re not shopping for anything else.
And so, their comparison-shopping, us against the other open source components that they’ve come across, and it’s kind of a best-of-breed sale, and that sort of sales process and a sale cycle is very well-structured. The requests are pretty standard, and we have lots of ways to service that with our sales, and our solution architects, and our partners.
The other type of customers that come to us are customers that have broad business problems.
They know that they want to build all kinds of services in the future, they know they need a broad integration backbone, they don’t necessarily know what the best architecture is yet. And they’re looking for us on a consultative basis.
In that situation, we spend a lot of time, with either our partners, or that account, doing solution design, putting hypotheses together, and it’s almost like an advertising agency, where you have to say, “Here are five or six different ways you could approach these broad class of problems, and here’s all the different configuration of, either our technologies, or other technologies out there that would let you deliver on that.
In that situation, it’s a longer-term sale, it’s highly consultative. And if you do a good job on the consultative side, then they’re usually more willing to pick our technology stack, get into a subscription with us, and sometimes even turns into a strategic relationship.
Michael Schwartz: It seems, like, for anybody in the enterprise integration business that channel partners are going to be important, but I’m wondering, are there other types of partnerships that have also been important for WSO2 to grow the business?
Tyler Jewell: There are really essential partnerships that happen in the open source realm in the ecosystem itself. Everybody’s competitive to each other, and everybody is complementary to each other now, and so it’s really important to build key technology relationships with other vendors, as a way to demonstrate confidence in the software that you’re building there.
I think that WSO2, historically, had kind of a very vertical stack technology that it has built from the ground up. It hadn’t done as much of that as some other companies had, but over the past couple years, we’ve been shifting towards more of that. And there are important relationships that we’ve got budding with vendors like Red Hat and Pivotal, which are, on one hand, both complementary to us, and competitive at the same time.
Those are pretty essential when we call those technology partnerships to go along side of the channel partnerships that we got.
WSO2 In The Next 20 Years?
Michael Schwartz:You have a 10-year product patch lifecycle, which is really long. How far ahead do you look? Do you look 20 years ahead in the future? Or, another way I could ask a question is, where do you see WSO2 in the next 20 years?
Tyler Jewell: Oh, well, you know, what I will say to that is, the company’s raised $45 million over these years, and for the first 12 years of its life, it was not profitable and cash flow positive.
So, I think that when a company is going through that sort of phase, it’s searching out for product-market fit, and when you’re looking for product market fit, and you’ve got a cash runway, you’re kind of forced into a horizon, which is somewhat tailored to your cash position along the way, although really good founders will look way beyond that.
With the company now cash flow positive and profitable, it completely changes our horizon, and I think we really think of things, and probably we like to think of them as 10-year horizons, although the pace of technology is moving so fast that 10 years just really, you know, kind of shock the hell out of me.
But we certainly are making bets that I’d be surprised if we don’t see any revenue. It could take us three to five years before we see revenue on some of those bets.
And it’s really refreshing to be in a position, where you can take those sorts of risks on as a company, while still supporting your existing customers. And for us, I think the biggest shift is, all of integration over the past 14 years has been essentially done, center of excellence, a big, ION middleware. And that’s even our classic stack, but that’s shifting towards a completely distributed, decentralized mode, driven by containers and kubernetes.
So, we’re just kind of over the moon tickled because we get to take all this domain expertise that we’ve got, but we’re basically throwing a lot of stuff out the door, and say, “If we had to re-envision all this stuff from a developer first code, first decentralized, first point of view, how would you make it all work?” and you just approach things very differently.
We’ve got about a hundred employees, working on things that I would say related to this sort of thing, and I really don’t anticipate seeing revenues from any of those things before ‘20/21. It’s that sort of rising that you got to be playing with.
Michael Schwartz: My last question is, any advice for new entrepreneurs about launching a business around open source software?
Tyler Jewell: Well, if you’ve already made the decision to go open source, and now you’re creating a business around that, be reflective of what are the principles and values for why you did open source the first place.
You need to hold to that and find a business model that allows you to not compromise on those elements – I think that’s really important here.
The other thing is to recognize that, take it one step at a time. Businesses are not built overnight, they’re built over a very long horizon, and you just need to have a mindset of just keep taking one step forward. And as long as you keep doing that, things will work out.
Michael Schwartz: Tyler, thank you so much for sharing your thoughts and experiences today.
Tyler Jewell: This was a real pleasure for me, Mike, thank you for having me.
Michael Schwartz: Transcription and episode audio can be found on opensourceunderdogs.com
Music from Broke For Free and Chris Zabriskie.
Audio editing by Ines.
Production assistance and transcription by Natalie Lowe. Operational Support from William Lowe.
Follow us on Twitter, our handle is @fosspodcast.
Tune in next week for an interview with Ofer Bengal, CEO of Redis.
Until then, thanks for listening.