Clément Salaün: Whenever you start to feel like maybe you're not going to sleep well tonight after having shipped this feature, that's a good starting point to start expanding your horizon in terms of what tools can you start to think about, what ledgers, and start your journey to that world. Eric Anderson: This is Contributor, a podcast telling the stories behind the best open source projects and the communities that make them. I'm Eric Anderson. I'm joined today by Clem, co-founder of Formance, who's talking to us about the open source project, Formance. Thanks for coming, Clem. Clément Salaün: Cool. Yeah, thanks for having us, Eric. Eric Anderson: What exactly is Formance? And we don't have to go into super detail, but what's the kind of one-liner for folks? Clément Salaün: You can describe Formance as a tool to build machine that move money. That's mostly what our customers do. And we try to focus on equipping you with this capacity to build all kind of things, so hence the quite horizontal statements that let's get into the details as we talk. Eric Anderson: And how long have you been moving money, Clem? Has this kind of been a life's work for you or? Clément Salaün: Yeah, good question. Yeah, I guess you can say this way indeed. So I've been in industry for more than 10 years now, and for some reason I was fortunate enough to always have been writing software that happens to be moving money in all kind of different ways. Back in the days, my very first project and way to learn programming was with a system from Minecraft server owners to sell items on their servers. So that was way back with the first version of Minecraft. So that's what initially got me into programming, and also into this concept of there's an inventory, there's money flying around, you don't want to mess up, you want to deliver the item to the person, and all that, and you want to capture the payment. So just got started in the industry of writing software that happened to be adding to section of funds, money movements, money payments and all the concepts that were quite critical. So that was back in the day and then actually started my career as a software engineer and took it from here, and always ended up working on a project that involve payments intensive features in a sense. So grew from there, worked for different kind of companies, marketplaces, financial institutions, and yeah, essentially that is how everything started from my own experience of having to be the software engineer in charge of making sure we don't double pay someone for example. That's how things got started three years ago when we created the company. It was built on the foundation that I had myself as a software engineer. I see in them a team builder and payments and payments operations. Eric Anderson: When I look at the open source, I see there's a bunch of repositories, and I think the website kind of advertises that you're a modular kind of system. So is there kind of one open source project or is this kind a collection of components? Clément Salaün: Yeah, for sure. That's fantastic question. And actually yes, I think what we're really aiming for here is to be akin to Amazon Web Services but for financial infrastructure. So we mostly started with our ledger repository, which is a core ledger product for platform financial institutions, FinTech marketplaces to store things like users balances for example. So this is how things started and this is the first historical product that we developed, but we briefly wanted to have an approach that is as mentioned like super modular in the sense that we mostly help companies create bespoke [inaudible 00:03:57] fans and there's no one way to do it. There's no single similar backend that they all use. There's no single financial institution that they work with. It's mostly a collection of different highly specific configurations. So to my opinion, it is not possible to address the problem. It's a very frontal approach and coming with a single product that does everything. And what we aim for, also from a product perspective, and also in the way we want to continue to engage with the open source community is to create these tiny products that you can use either in a standalone way. For example, we just extracted our transaction DSL from our ledger to a single repository in order to let you use it without having to use our ledger as well. So that's the approach that we currently push, and I think it's much more beneficial bus to us, our customer, but also the OSS community that will be consuming and interacting and then participating in the development of those projects. So that's the ambition here on what we are doing today. Eric Anderson: So if you're the kind of AWS of moving money, is the ledger kind of the S3 or the EC2, the foundational product or? Clément Salaün: Yeah, definitely. Yeah, that's a question we often have, like what's the relation between the different parts. So the ledger could probably be seen as the either the EC2 or the RDS, depend on how you say or you view things. It's probably going to be, if we describe Formance as a computer for moving money, the ledger is definitely the storage layer, but we also take care of compute and network, we think like connectivity to financial institutions. So yeah, let's say the ledger is the storage of that computer and let's say it's either the RDS or S3. That would be pretty close. Eric Anderson: And at some point I want to talk to you about, and maybe this is as good time as any, the decision to go open source. I can imagine if you're building a developer tool, it's maybe more natural to consider that, but it does seem like going open source has been kind of core to your value proposition. At what point did you decide that? What was kind of the story behind it? Clément Salaün: Yeah, definitely. We decided to go open source from the very start of the project. As you mentioned, to be perceived as a developer tool was indeed part of the initial version we wanted to develop. So being open source was a no-brainer for that reason. I think probably all companies and all OSS projects have to find their flavor of being open source, and for us, I think what's interesting with open source that it's both about getting some traction and some users from the OSS community, but it's also about providing trust. We tend to work a lot with enterprise scale companies and they need of our free insurance relative to the own ability of the products that they are going to be using, especially for the most foundational component of their products, so anything touching money. So for us, the trust component of OSS was really important. That was sort of a wild guess that I had at the start, but that proved to be right over time. We knew that we explicitly won some contracts by being open source that we just couldn't have won otherwise. Eric Anderson: Maybe to help us make the product more concrete, what types of companies use Formance? These banks who have much of this built but they're kind of revamping it or is it FinTech companies? Clément Salaün: I think we're probably in the age of these labels like banks or FinTech are becoming super blurry in the sense that we tend to think more in terms of activities instead of categories of companies. Essentially we have some customers that started as an ERP, for example, for a specific customer vertical, so providing them their day-to-day operational software. But then that grew with the FinTech parts, essentially embedding payments into the ERP and starting to provide loans to their ERP users, and that are now adding a marketplace component to the ERP as well. So we tend to think more in terms of activities. We do sell to product first software company that have a high intensity in the level of sophistication in terms of financial features that they're adding to the system. So yeah, we have some companies that work with us that looks like bank in a sense, and we have some companies that are a bit more further than that, that look closer to traditional software, but that happen to be starting to embed banking features, for example, or starting to embed a marketplace component in their software. So it's definitely getting buried and we couldn't be more excited by all this long tail of crazy use cases we see, and to double down on the analogy here on these huge machines that our users are building, moving monies in all kinds of directions as part of their product. Eric Anderson: So just so listeners who are building products can map their needs to what Formance might do, if they come across a situation where they're starting to feel like they need a ledger, they're starting to do transactions that are bank-like against an account, that's when they might reach for something from Formance. What else would be an indicator that like, "Hey, I'm doing something that Formance can help me with?" Clément Salaün: Yeah, that's a very good analogy. Anytime you're [inaudible 00:09:25] things like storing a user balance, starting to do things like, "Hey, Eric's balance on my platform is like a 100 USD." Anytime you start instructing different payment processors or just financial partners like a bank to execute a wire, you will start to enter into all kind of consideration about the code that you are writing will probably need to leave in a special place. If you don't take care of it and start to think about that code in a different way, as the codes that will for example create the algorithmic feed for your system or things like that, you need to isolate that product, the code that touches money in a very special place because that is the only part of your product that has some kind of asymmetric exposure to risk. It's okay to have likes counter. It's not going to be okay if you wire the wrong amount to Eric. So yeah, whenever your thoughts to feel like maybe you're not going to sleep well tonight after having shipped this feature, that touches on the money your Eric, that's a good starting point to start expanding your horizon in terms of what tools can you start to think about, what are ledgers and start your journey into that world. Eric Anderson: I don't know if you're in Paris today, but I think the company is largely based in Paris, and the banking systems across the world vary. Are you kind of relevant in all geographies? Do you focus on any kind of regulatory framework more than another? Clément Salaün: Yeah, that's a good question. So we tend to see ourselves as default global in the sense that our customers start to be default global as well. We have customers either in Europe or in the US or in LATAM as well and in Asia, and they all have the ambition to go cross countries and to have to deal with local specifics at the end of the day, but still maintain their business logic at a default global sort of definition container. So yeah, we sort of try to unify the core of your system on four months regardless of local specifics, and then depending on where you go in the product and in the Formance platform, if you start using things like our connectivity system to banks and payment processors, you will indeed start to involve different partners, like depending on whether you're going to be using ACH in the US or whether you're going to be using SEPA Instant credit in Europe. That's going to be a different patchwork of partner that you will work on, but we really focus on agnosticity. I think that's quite important. Historically, software that we're providing capabilities to handle money in a sense tended to be glued to a specific set of partner like NJPMC for processing payments or having bank accounts in another processor for a specific country. We want to break that bundle and to provide you with a software so that you can bring in the best partner to work with you locally and across all the countries you're going to be operating in. Eric Anderson: One more question on being in Europe, there's a lot of great open source happening in Europe. Why is that? Part of me wonders if the world of open source has no geography. If you're in open source, no one cares where you are, and so it gives you this global presence, but what's your take on why the great open source is coming out of Europe today? Clément Salaün: Yeah, I'm not sure I have a definite answer. That's definitely something we can see on the markets. I definitely think that you're right on the no boundaries. What we see in crypto is probably good overview of what everyone is trying to achieve in the different RSS communities. Regardless you're in Berlin or in New York, you can produce the same code and we try to unify everything at the global level. So that's not a surprise to me that there is a wave of RSS companies coming from EU. Depending on your project, like for us, for example, we're still going to be opening an office in New York in the US in Q1 '25. So we feel that it's still good to have a local presence everywhere, even if you operate mentally in a default global stage. But yeah, I'm as interested as you in the answer to that question. I am afraid I don't have definite answer to this one. Eric Anderson: And there's a lot of talk today about switching gears a bit to talking more about open source business models and how people monetize. Was that clear to you for your company at the beginning, or is that something that's evolved over time, and how are you approaching building a business on the open source? Clément Salaün: So essentially for us, we had the ambition to develop a strong OSS monetization model, and we had this idea that it was going to be possible. At the very beginning of the company we weren't sure about the details, but we figured them out as we continue to onboard customers. Same as the open source flavor you want to create for your company and write on for your company, I think you have to think about your specific OSS monetization model. For us, it started to be a success in term of clarity of the pricing and comfort of the customers with the pricing when we started to think into three dimensions such as hosting different premium features and then support dimension. So we found our unique equation. I'm afraid that this unique equation is not going to be working for all the auditors here and that they will need to do the work themselves as well, but there's definitely a bit of literature you can take some cues from. I think we also had another advantage that is to be factored in here is that when you work on business-critical money-moving software, there's no strong opposition to the idea of paying for the software. Companies tend to come with the idea of yeah, there's going to be something that basically operates the whole company and that's everything depends on, so of course we want a commercial relationship, and it's good that it's OSS, but we are still going to be wanting to pay this company for a privileged point of contact and ability to know that we can confidently operate the software. So that's another advantage that we had in the payment space that is probably going to be playing out differently in another field. Eric Anderson: Yeah, taking those open source comparisons further, you have a surprising, at least surprising to me maybe, large active community on Slack. You could take community to GitHub star ratios, and I think yours is the highest I've seen. And maybe that's, as you kind of mentioned, if you were a JavaScript framework, there's kind of broad appeal, but shallow engagement, you seem to have narrow appeal to people building these particular applications, but deep engagement. How would you characterize the community? Clément Salaün: Yeah, yeah, definitely. Yeah, I can only echo that we were surprised as well at this level of engagement on the Slack, for example, versus the ratio to star. We weren't sure about what this ratio is going to be at first, but we are indeed able to initiate those conversation pretty easily, I think, thanks to the fact that as we briefly mentioned at the beginning of the podcast, those topics, you might end up in a position as a developer where you start writing some codes that is triggering a wire and you're like suddenly, "Oh, shit, how can I do that properly?" So you want to reach out to other folks who can also tell you more about the best practices and just be a safe advisor to whatever you're doing and so that you can get another proof from someone that yeah, you are actually thinking this right. And you can ask questions about not only the technology and the specific products, but also about writing software that moves money. That's a first, I would say, part of contact that we get from few users today. Eric Anderson: One of the risks I see in open source founders is that they end up having two audiences. One is the people who get excited about the open source, and the other are the customers, and there's not a lot of overlap, and so they have to weigh, should we build this feature that the community wants or should we build this feature that the customers want? I'm sensing from you that you don't have that tension, that people who are engaging with the open source are very aligned with the people you expect to sell to. Clément Salaün: Yeah, we are fortunate as well in that sense. I think it's definitely tied to the ratio of stars to engagement, as well as if you're writing this type of software, you're likely to have some commercial needs around it, because whenever you're building a FinTech or marketplace and you're actively starting to move money for real, it's not just a site project that is basically an experimentation. You can probably get that kind of audience if you create a JavaScript framework, but if you're actually providing solution to people trying to make a wire, there is likely an actual project, an actual body in here. So I think that that narrows down the audience, and you're absolutely correct, we don't sense a lot of tension on that front today for sure. Eric Anderson: I think you mentioned the word crypto before I did, which I was hoping would happen, but now that you mention it, I'm sure you get questions all the time about is this Web3 ledger, is this crypto blockchain under the hood, and why or why not? Clément Salaün: Yeah, crypto is definitely an inspiration in the sense that on crypto technology, money really is natively programmable. There's no blocker to moving money around and to whatever you can do with money, but in traditional financial infrastructure, programmability is more like a goal that is on paper, but that in the real day implementation is something that you constantly have to put some effort to, to make happen because you are interacting with many institutions with different APIs, different way of [inaudible 00:19:20] risk. Nothing is quite deterministic in the world of traditional financial infrastructure. So we're inspired by the quality of programmability that you get in crypto, and try to infuse our products with this quality of programmability in different layers. Like for example, Numscript or transaction DSL is one example. We developed a language to describe financial transaction and to make them happen in your system. We felt that that was missing part in terms of the standard typical tools that people were used to work with. So we wanted to bring that side from crypto, and then obviously comes the question of what is the frontier between crypto infra versus traditional infra? And I think today more than ever, they're definitely being blurred a lot, especially with stable coins and users that we see creating hybrid setup with a banking partner for Europe, for example, but in the US use stable coins instead. So I'm really excited about this side of things, but yeah, it's also why the word crypto is likely to come up quite often and more and more. If we were to do a podcast next year, I think probably 50% of it would be about crypto instead of 10%. Eric Anderson: I'm a little surprised this hasn't emerged already. What were your kind of thoughts as you looked at this market? Fintech has kind of been a bigger and bigger deal over the last, is it a decade? Are there other solutions that people run into or come from? Is Formance the first to do this, I guess? Clément Salaün: I think it just takes time, because I mean last decade we were still figuring out how to handle payments on the internet roughly, and it takes time before you get this kind of iteration function to work in financial infrastructure and FinTech. Obviously things can move fast in social networks because there was no inherent blocker to the speed at which the clock can tick in term of iteration. But in term of FinTech, we still have to bump into issues with the regulator who has to disagree with something after a few months, and then that changes the next wave of what we can do over and over again. So the clocks move slightly slower in FinTech, even if it ends with the suffix tech. It still starts with fin, so the clock time is still a bit different, and the clock speed moves a little bit slower, but still it's moving steadily and I think there's no surprise that solutions like Formance starts to appear, especially as companies start to not only, as we briefly mentioned as well, not only is it stop being labeled as a specific category, but starts to be creating all kind of activities such as an ERP, embedding a marketplace and a landing program. We're seeing that hybridization a lot today. That was probably not the case last decade. So yeah, I think this makes the case for this type of tooling to emerge today with this complexification of everything and hybridization of use cases that we can see now that we have enough feedbacks from the regulators, we know what works, we know where it can go, and people just start to come up with new ideas. Eric Anderson: You do have a lot on the website about the various ways to deploy this. You have a fully hosted cloud, this CloudPrem one and two, these are kind of versions of hybrid that suit different people, is that right? Clément Salaün: Definitely, yeah. Yeah. Again, I think this ties to the nature of the product, like some OSS product will be more or less sensitive to the ability to run on prem and the ability to be run into under tight control from the company running it. In our case, our customers are going to be justifying their technology stack to regulators for example. So it was a no-brainer to provide those options. We invested quite a lot in the CloudPrem model that you mentioned, so I think there's a thing that works well for us. So we basically created a Kubernetes operator to try to live in the best of both worlds, but still give a cloud-native experience that still have some level of control relative to how customers can deploy the thing on their own system. That's the thing I discuss sometimes as the OSS company's founders. It's not easy to keep some level of control relative to how your thing will operate. We sort of achieved that with the CloudPrem you mentioned, which is basically a Kube's operator that drives the whole deployment end to end, and tries to live in the best of both world of being cloud native but still self-hosted and with minimal overhead in terms of what you have to do manually to deploy the thing. Eric Anderson: Clem, I'm interested if you've developed any kind of unique ways of engaging with the community over at Formance. I think every open source startup has to wrestle with like, okay, who's going to monitor the Slack group and engage there? Do we do events? Do we do content? What have the things that you figured out at Formance that you feel like have become key to your growth? Clément Salaün: Yeah, definitely. I think we're going through this cycle as well ourselves of figuring out what's the future for the Slack, what's the future for the website? How do we continue to help people learn about FinTech, about financial infrastructure? What's been working well so far is content and pieces that we can write about our convictions relative to how you can build a ledger, for example. So we'll [inaudible 00:24:50] double down on that. On the community front, we are taking cues from what some other companies are doing. I think we are highly inspired by POSTHOC, for example, who I think is starting to nail down this concept of community under a single tab on their website. So we still have much iteration ahead on that, but it's as monetization flavor. You have to find your own version of what works for you. For us, this is likely to be propelled by the fact that engineers are actively trying to understand the thing. It's very learning-based and content-based essentially. Eric Anderson: Yeah, POSTHOC is one you think does this well, is what you're saying? Clément Salaün: Yeah, definitely. If you go to their website, like this sort of crammed into a single tab like this concept of community with a forum with the blog posts and everything. And I think we saw them iterate as well, like having a Slack first where people started to ask questions, having a separate blog, having the docs. I think has already started to nail this down with this unified version of what they call the community. So yeah, to the audience here, probably some interesting cues to take from there as well. Eric Anderson: Help us understand the state of Formance today. What does the future look like? To what extent are you kind of beta GA and you have these separate kind of product modules? Do you have all the ones you kind of want or how complete is the platform? Clément Salaün: I think in terms of life cycle, each part of the products gets the opportunity to have its own life cycle. We're now in a mode where we moved off from a monolithic product to highly modular product, so the ledger can have its own life cycle regardless of whatever we are doing and the connectivity to banks part of the system for example. So each part can have its own life cycle, and we have five modules today, five features delivering modules today, but this is essentially just the beginning. We are still quite a lean team of about 18 people and we'll continue to expand finding to be about 50 people on the team developing. What we have on the roadmap is about 30 modules. So we really want to go really fine-grained and always at all times address things in a way that let our users combine the different modules of the platform. If you have a specific problem on your reconciliation issues, you shouldn't be cornered into something that is a full-blown product that will be hard to configure. We want to give you small gears that you can assemble yourself, hence the ambition to continue to generalize. And for example, the ledger is going to be split in three different modules, and I will continue to expand into other financial operations areas as well. So yeah, we just onboarded some fresh funding to take on that ambition. So if you follow us some GitHub, you will likely see modules popping up every now and then at the upcoming year. So we are really excited by this. Eric Anderson: So you had a single monolith and then you broke it up, particularly you pulled the ledger out kind of on its own, and now you've got five modules, I think you said. You have a vision for 30, and in particular the ledger as an example is going to be broken up into three you said? Clément Salaün: Yeah, precisely. Eric Anderson: GitHub doesn't lend itself to a good marketing of a family of projects. It's kind of like a single project paradigm. How do you kind of position an open source project that involves lots of repos? Clément Salaün: To be honest, I think I still have PTSD from these kind of discussions with the team, but how do we optimize for stars versus how do we optimize for the actual ability for users to take a small part of the product so we learn in on this model where each project, each product is going to be living in its own separate space. It's true that GitHub favors big repos today, where people can just go to company/company in terms of what is the repo and just start the project. It will be different for us. We accepted that cost, so to speak, in the sense that it favors the just ability to adopt the product well without committing yourself to a big thing. So that's the [inaudible 00:28:58] cost you how to pay in terms of ability to have one big repo with [inaudible 00:29:02] stars. But on the other hand, it's also helped. Like we started to launch some of these projects in a standalone way, and even if it's harder to collect a large store account, I think it feels better when you launch something in the sense that for users discovering the launch on Hacker News for example or whatnot, they just have to understand one precise thing. They don't have to understand Formance as a whole. We are not launching Formance every month. We're launching specific tools that can solve a specific problem for them like Numscript or Transaction DSL. You just have to understand that there's some saying to craft transactions. You don't have to understand our banking connectivity layer, and it's much easier to do atomic standalone launches with that approach. It's definitely a double-edged sword or double-sided coin, if I keep the terminology in the FinTech space. Eric Anderson: Yeah, no, you're right. Maybe you're sub-optimized for GitHub trending, but you get a better user experience and then you get to ship more frequently, and you celebrate those launches, even have sub-communities. Can you imagine a world where you have sub-communities where people are like, "I'm really actually mostly interested in the ledger or something and less the other stuff?" Clément Salaün: Yeah, I think it started to be the case. It's still the early days of that, but we start to see people just forking one part of the system and forking Numscript for example, to add something about adding web assembly to it and stuff, and they couldn't care less about some other parts, so that's great. But they still congregate in the same spots for now. So on the community Slack and just a Formance GitHub as a whole, so that's something you have to continue to make it work in the future as we, especially land [inaudible 00:30:51] this different repos that will have their own life cycle and store accounts. Eric Anderson: Numscript seems to have gotten some attention on its own. It seemed to be trendy on Hacker News once. Is that kind of core to Formance? Clément Salaün: Yeah, definitely. That's something we extracted from the Ledger, which is essentially the DSL that you can use to craft your financial transaction. And we figured that some companies really fell in love with that, but for internal reasons had some complex deployment path to migrating their whole core ledger because maybe it's invested years into building that core ledger. And I think it's much easier from a user experience just be able to use the parts that you own. So if you fall in love with NumScript and the way to have this declarative model to craft your transaction, we're interested in helping you do that. We don't want to force-feed you this whole thing. We want you to use the parts that is realistically deployable on your system. So yeah, NumScript is a good example. So that used to be core is now satellite, but still belongs to the Formance ecosystem for sure. Eric Anderson: Great. Clem, I think I've covered most of what I wanted to cover. Anything I missed out, good stories that we should include or things you wanted to cover? Clément Salaün: We could probably do a whole podcast episode about the horror stories that we've seen in Fintech and in the financial infrastructure, like crazy scenarios where companies end up in a stage where they just lost like 500K at the end of the month like we- Eric Anderson: Oh, these are bugs that cost real money. Clément Salaün: Precisely. Yeah. These are things we see every month and these makes for the most interesting stories. They usually are attached to specific company names or they're hard to tell. It's really easy to mess up in that role. And we constantly talk to companies every month who lost this kind of sums or just cannot track it, or not even sure that they lost it. That's the crazy stories that we can see in our world. Eric Anderson: Yeah, I would imagine if you knew you had misplaced money, you would just fix it, but the real concern would be a year ago you realized that a bug's been around for a year and we may have been misallocating money this whole time and nobody noticed. Clément Salaün: Yeah, precisely. And you can end up in a situation even where things sort of even out in a way so that the problem remains hidden because you overshoot 100, but you also miss 100 in here. And so everything from a summarized point of view, it looks sort of correct or at least off by not that much. And then you discover, as you mentioned, precisely that after a year that things are off and those makes up for good movies, I guess. Yeah. Eric Anderson: So I was in Paris last year, almost a year ago for KubeCon, and I learned all about the education system, big math focus. Clem, are you kind of a disciple of the French math system or education system? Clément Salaün: I can probably relate partially only to that topic. I did an architecture school myself, so quite unrelated. But yeah, I can see quite talented folks around me. I think this engineering culture is definitely present in France with these top schools that people attend, and I think we have a good foundation to continue to scale like technology companies in here if we focus on the engineering talents that we can leverage. Marketing and product talents are another story in here, but for the math part, I think we get the math right in France and Europe. Eric Anderson: Well, thank you for not only coming on the show, but for the gift of the open source of Formance. There's probably a bunch of products being built today that maybe either wouldn't be built or are being built much better or quicker. As we wrap up here, tell the folks where they can get involved if they want to. Clément Salaün: Yeah, for sure. I think just taking out the GitHub with the issue of it not being monolithic as we discussed is a good starting point. You can go over some of our latest blog articles. We like to geek out about how data should be structured on ledgering systems and things like that, and about entangling bits of money. So if you want to geek out on this type of content, feel free to check out our blog and then joining the Slack is a good next step to start to initiate some discussions with us. Eric Anderson: Thanks, Clem. Clément Salaün: Thank you, Eric. Eric Anderson: You can subscribe to the podcast and check out our community Slack and newsletter at Contributor.fyi. If you like the show, please leave a rating and review on Apple Podcasts, Spotify, or wherever you get your podcasts. Until next time, I'm Eric Anderson and this has been Contributor.