Mike: Hello, and welcome to Open Source Underdogs! I’m your host Mike Schwartz, and this is episode 62 with Amandine Le Pape, Co-Founder and COO of Element, the software startup behind Matrix, an open standard for secure, decentralized real-time communication, which you can learn more about at Matrix.org.
This episode was recorded in early February at the inaugural State of Open Conference or SoCon, which was held in London at the QEII Center in Parliament Square.
SoCon was made possible by the inexorable tenacity of Amanda Brock, who leads an organization called OpenUK and is the editor of the 2022 Oxford University press book, Open Source, Law, Policy and Practice, (2nd edition).
If you’re an open-source founder and you haven’t read this book, go to Amazon right now and order it. If you can travel to London next year, I hope to see you at SoCon 2024. If it’s half as fabulous as this year’s event, you definitely should not miss it. I guess that’s enough gushing about SoCon, let’s get on with the interview with Amandine.
Origin of Idea
Michael: Amandine, thank you so much for joining us.
Amandine: Thank you for having me.
Michael: As one of the founders, I have to ask what’s the origin story of Element? And at what point did you think about starting Matrix, the open-source repository and Matrix of protocol?
Amandine: So, it goes back to almost 9 years now, in 2014, where we were a team selling commercial messaging apps to Telco’s, incubated in a big corporation called Amdocs. After a while, we were really annoyed by the fact the whatever we did with our apps, WhatsApp would always win. WhatsApp would always be on the front page of the Telco’s website, next to the app we were building for them.
And this fragmentation even from a user perspective is really bad. For email, I can talk to anyone with my phone, and can send SMS and call people whatever phone they use, whatever network they are on, while on chat I have to actually install a new app every time someone wants to use a new one. So, we came to Amdocs and proposed them to actually try to fix this by creating an open standard for communication.
Having built on top of all the other existing standards before, we had learned a lot and thought that with the professional team, we would be able to truly bring something to the world, which was able to answer the needs that we had of interoperability for chat and messaging voice over IP.
Origin of Company / Project
Michael: That sort of answers my question, but you’re working at Amdocs and you’re thinking, “Well, wouldn’t it be great if we could federate these chat servers,”, but how does that lead to actually starting an open-source project and the founding of the company?
Amandine: So, what we do actually is create the open standard and open protocol – we set it up as an open-source project. So, within Amdocs, we set up completely something independent with a new brand, completely open-source. And we started working on it for three years, actually building the reference implementations, defining the spec, etc.
After three years, there were other companies building on top of Matrix and monetizing it, like Erickson for example, selling communication systems for banks in Sweden. Whilst the core team, we were still building the project and not funding it.
So, based on that, we thought that it would be better to actually set up an independent company and try to monetize Matrix on the side. So, that’s when we actually spent out of Amdocs, set up Element as the commercial company, which builds the flagship app itself – called Element – and selling services and proprietary product on top of Matrix. And we also set up the Matrix Foundation as an independent nonprofit, which is the custodian of the standard itself.
Michael: The Matrix Foundation governs the protocol, but it also has some software in there. There’s a reference implementation, I think, in Python. Who contributes the code to that software project? Is it mostly Element or, are there community contributions?
Amandine: Basically, Element contributes a lot of code towards reference implementations of servers and SDKs and some of the application services. And it’s literally donating the IP to the foundation, so that it’s completely independent from the commercial entity. But beyond that, Element contributes the reference implementation server Synapse, we also have another server called Dendrite – all the various SDKs for iOS, Android, etc.
But then, the community, on the other side, is actually building their own projects and building their own reference implementation servers, their own SDKs, Bridges, etc.
So, whilst Element contributes a lot to some of these projects, which are the ones which are often used in production by governments or enterprises. The community is very vibrant, and I think we’re close to 5,000 contributors across the entire Matrix project and the different repos.
Michael: So, there are a number of communications platforms out there, we’ve actually had Ian Ten from Mattermost and Gabriel Engel from Rocket. And there’s Slack, of course, is out there. What makes the Element commercial offering different, and why do we need another communication platform?
Amandine: So, if you’re just trying to have your own communication system that you can run yourself, then yes, you may want to just go to Mattermost or Rocket.chat. The difference with Element is the fact it’s built on the open standard – it means that once you run your Element and Matrix deployment, you can communicate to any other deployments out there. The same way that when you deploy your mail server, it federates with all the other email servers out there.
And actually, Rocket.chat built a bridge to Matrix, so now, Rocket.chat can communicate with Element without a problem. And the interesting thing is that was under the impression of the Swedish government, which brought all its vendors into one room and said, “Guys, we cannot use SaaS solutions for communications – we don’t want to designate one vendor for all the government, ministries should be able to choose the app they want to use. So, you have to federate, you have to interoperate – figure it out, find something.”
By the way, there is this standard called Matrix that should do the job, but that means that today, if you use Rocket.chat, you are able to participate in the Matrix network. So, the interesting thing is also that when you use Matrix for your communication, as for a government for example, it really solves the political problem.
Because different ministries will own their own communication, they will deploy their own servers, they run it wherever they want, under the security they want, but yet, they are able to participate in the wider network and the ownership of the discussion is shared between the two. So, you completely skip over the “Shall we use your system or my system” to come to work together on this project.
Building Matrix v. Element Brand?
Michael: So, here at the conference, I noticed that there’s a Matrix booth, but there’s no Element booth. Why did you feel that it was the right thing to do to promote Matrix and not the commercial company here?
Amandine: It felt more aligned with the idea of OpenUK in terms of building open source in the community side of things. Basically, the same way, this weekend we had the Matrix. It’s Matrix which is being most represented – we don’t even have an Element sticker or an Element banner out there. Even if Element is open source too, but really trying to grow the ecosystem in the community for Matrix.
Michael: This is a little bit of ancient history – but do you remember a platform called Diaspora? Were you thinking about Diaspora at all when you started Matrix?
Amandine: So, the thing is that we’re addressing very much the social network side of things whilst we wanted to go for the full communication, real-time communication system. In terms of whether we thought about them when we chose the decentralized structure, actually it’s more inspired from Git, and basically the idea of replicating the content of the conversation and the history of the conversation on every server.
Michael: Let’s talk a little more about the community. What areas do you think the community makes the most contributions? Is it in Bridges, Bots or Widgets, which is, like how you extend the functionality? Or is it in the core server? Where do you see the most activity from the community?
Amandine: So, the way we built Matrix is to make sure that the development of clients was super easy. Because what we saw in the past is every time people wanting to add chat in their communication systems, everyone thinks it’s easy – it’s just sending a message. And then, if you want the actual features of history, etc., it’s tough. So, we wanted to provide an easy way for any web developer to add chat to their system.
So, all the complicated stuff is in the server with a very simple client-server API, just HTTP/API that anyone can use. It means that we have a lot of contribution for clients, a ton of clients out there – all sorts of clients, from common line to mobile, etc. We even have clients running on Nintendo DS for example. And a lot of Bridges as well because people want to connect to existing networks out there, so they tend to build their own Bridges. There is also contribution to the core, to Synapse itself, and we actually have the rest SDK as a new core project, which is led a lot by the community. I think it’s a quarter or a third of the developers actually are not employed by Element. And it’s like the core contributors on this project, and it doesn’t even have an internal room -everything is really in the open for the rest SDK.
Michael: So, Element is very transparent about the monetization strategy. It’s fairly simple, it’s based on monthly active users, and it looks like it’s the same price whether you host it yourself or whether it’s SaaS – can you talk a little bit about the logic behind the pricing model? How did you get here and were there other pricing models? And did you get it right the first time?
Amandine: All those are complicated. So, when we started, we thought that the easiest way to get going would be to provide it as a SaaS platform, because it’s end-to-end encrypted and based on an open standard. So, even if we run it for people, it means we don’t see what’s inside, what’s happening on the platform due to their encryption, and people can take the data and move it to their own server later on if they want.
The fact is, our customers actually want to self-host – be it on-premise or in a cloud – so, we had to basically make sure that the on-premise hosting was actually served very well. That’s why we ended up, this is — I don’t know which alteration of pricing model it is, but it’s definitely not the first one. And yet, the idea was like, okay, let’s try to simplify one single price in the cloud or on-premise, it should be the same, because in the end it’s pretty much the same service.
I think we are about to release unless we released a couple of weeks ago – I can’t remember – a new pricing model, which is a bit like a more integrated basically. Here, we had, I think it was a three- or four-dollar user a month. And then, you can add add-ons on top of it. But we’re trying to build packages where basically the add-ons are included, but the total price may be higher, so have more of a split like this.
Michael: So, what does the sales motion look like? Do people download one of the open-source and find you? Is it more an inbound motion or, are you out there, trying to find like who wants a service – like what is the sales, how have you built the sales team?
Amandine: We hired our first salesperson three years after starting the company because everything was coming inbound. Basically, it’s very much the community being here and enthusiastic, and then, you have people who tried on personally or investigated, and then bring it in-house and suggest it to their governments or their enterprise as something that would be useful.
The community has very much been our main source of leads, and so many times you have a first customer called, and then you see one guy sitting at the back of the room with the Matrix hoodie, and it’s the one who doesn’t say anything. But you know it’s the guy that people actually listen to and trust because, from technical perspective, they are the expert. So, we now have a sales team, and we’ve focused a lot on making sure we have content and marketing, etc. do a bit of outbound, but yes, the fact that our tech is good and take his, “No, it is good.”, is our best-selling point and our visibility as Matrix as well.
Michael: Have you found that there’s some natural segmentation in the market, either vertical or by application, like how does it breakdown? It seems like it can be used by so many organizations, so does that drive your marketing department crazy?
Amandine: Yeah, it’s tough to focus when you can be used by anyone, basically you can solve pretty anyone’s problem. When we started, before setting up Element, we looked at the different business model to figure out how are we going to monetize it to actually fund the development. We ended up with 72 business models. We went for one which we thought would be pretty easy, which is enterprise collaboration and starting on SaaS, etc.
We looked at one of them was public sector communications, and we were like, “Huh, that’s a good match!” But oh, my God, no way we’re getting into this, it’s going to be endless sales cycle tenders – no way! Guess who knocked at the door? It was the French government! Because in the end, it’s such a good product market fit in terms of no vendor lock-in because it’s open-source and open standard, ability to run yourself, end-to-end encrypted, decentralized, so it matches the organization, and each ministry can have their own deployment.
So, we have really seen a good product market fit there as either generic collaboration tool, like French government for example uses very much as with team’s replacement, or also, for specific use-cases like defense. Because you can take Matrix into extreme environments, you can use it peer-to-peer, in a mesh network, you can use it low-bandwidth on ships. So, the US Navy is trialing it on ships. Their ship has their own server, and if your submarine is actually going down, loses contact with the network, they still can talk locally. And when they come up, then it merges the history, and they can get back in contact and see what has been happening with the rest of the fleet.
So, all sorts of very specific use-cases, which I’ve been working on. But overall, it’s very merged, we need sovereign communication, end-to-end encrypted, we need to replace WhatsApp because it’s not compliant and it’s centralized, and we need to replace teams because it’s not even encrypted and centralized.
Michael: You’re still a pretty young company. When was the company actually founded?
Amandine: Element was founded in 2017.
Michael: Okay. So, have you built out the partner network at all? And are partners helping you deliver or have partners become a distribution channel for you at all yet?
Amandine: We’re just starting. Literally, we’re still building up offers, contracts, etc., because we’re seeing a lot of companies who are actually helping institutions deploy their own NextCloud or LibreOffice and this sort of things. And now, they’re coming to this guy saying, “Can I have my Element deployment as well please?” So, there’s this sort of partnerships that we’re looking at, but it’s very, very much the beginning.
Michael: Maybe I can degress back into the tech, or maybe this is a tech / business question, but it seems to me like one of the challenges around a federated system would be trust. “Yes, I could connect to anybody, but if I’m a submarine, how do I know I’m not connecting to an enemy’s ship? Yes, I can connect to anyone, but how do you know who you can trust?”
Amandine: You can set up securable gateways, for example, which are going to give you an opportunity to apply business logic to it. Like, for example in the French deployment, you don’t necessarily want anyone to be able to message the top layers of the government. You want to put like civil servants in the government, can invite external people civilians into a discussion, but the other way around is not true. So, you have tools like this which says, “Oh, your server can connect to anyone, or, your server can only connect to this subset of domains,” and this sort of thing. So, we rely on additional tools around it to do that.
Michael: What are your plans to foster growth in the ecosystem? So, not just the open-source contributors, but also sort of other companies who maybe have a commercial interest in working with either the Matrix protocol or the core Matrix open-source project?
Amandine: The idea from the start was to create Matrix as big as the biggest ecosystem as possible. And Element being the leader of it, but maybe like ten percent market share, the same way as Google is probably the leader of the web market, but it’s like tiny market share in corporate reason – it’s just a very big market. We are trying to encourage companies to build on top of it and support them.
Within the foundation, we want to set up at some point the ability to provide grants. We’re setting up a membership model within the foundation and then the ability to actually be able to fund all this money to other people contributing to the ecosystem. So, that’s the kind of things were trying to do. But we are very, very conscious about making sure that others have the space to grow within the Matrix ecosystem, so that it’s basically rising up the sea for everyone.
Michael: One of the key differentiators for Element is this aspect of decentralization and federation between servers, but that’s not proprietary. Are you ever concerned that maybe your key differentiator is also open and maybe, overtime, not unique?
Amandine: So, the beauty of an open ecosystem is that people will complete a value, so it’s up to everyone to find where the best value lies and bring their own expertise. So far, we have been relying a lot on the fact that we’ve been the expert in Matrix, having created it. So, we are the best to run servers at scale and this sort of thing. So, people come to us. We are at the point where other companies are now competing with us at that level, providing hosting.
So, we need to bring our expertise into more specific use-cases, and maybe it’s this specific property product for use-cases like around security, and that sort of things that we can build proprietary, and we can license, because we know best how the protocol works and we can go around these things. So, yeah, it’s up to us to be creative.
Another thing is that the vision is that everyone would be communicating via Matrix in a few years from now. And some of them will be using Element, some of them would be using all their applications out there. But we do provide today as an Element’s kind of app store, where you can install an integration to GitHub or Bridge to Slack, etc. One idea is that this App Store would be available to any other applications out there and can become one of the main points of monetization potentially for Element in the long term.
Michael: On the product side, which product are you the most excited about do you think has the most growth potential? And were you’re investing – perhaps more in R&D?
Amandine: Right now, we are rewriting the Element application, especially on mobile, based on the rest SDKs. So far, we had an IOS and Android SDKs. We are rewriting it from scratch using a Swift and Kotlin for Android. Basically, trying to get to a point where Element is as performant and fast as Telegram, with a user interface which is super simple. Because that’s been our biggest problem so far. Element has been very slow and bloated, and that’s really hindered the growth of Matrix as a whole.
This is the thing where we are investing a lot, and it should be out in the public test flight for iOS next week or in the couple next few weeks. And it’s really exciting because it’s like you started with some of the biggest accounts, and it’s like 100 milliseconds start – it is really nice.
Europe tech startup challenges?
Michael: We’re here in Europe, and the founders are European – do you think that there are any challenges about being in Europe versus Silicon Valley, and what’s it been your experience of tech startup in Europe?
Amandine: Our market is very much around privacy data sovereignty and this sort of things – which is huge in Europe – so, somehow, it’s probably easier for us to be based in Europe because we have Germany, we have France, and they’re all so advanced on this angle of data sovereignty and privacy. I think that was really good for us. And it’s now a real challenge to cross the Atlantic and see on the other side if we can address this market as well. Because in the US, as far as I see it, the clouds seem to be pretty much the default and it’s less questioned than here.
Advice for entrepreneurs?
Michael: So, every startup is kind of a marathon and an emotional roller-coaster – do you have any advice for entrepreneurs, especially entrepreneurs who are looking to use open source as part of their business model?
Amandine: We just had an interesting panel around the community side of things, so one of my advices would be if you are actually wanting to do it, think about why. If you want to do it right, really take care of your community. Because the open-source company without a community, it’s missing the point of doing the open source. So, make sure you manage to grow a nice community because this will help you so much. It really makes the experience something different, and you know, otherwise, it’s just usual tech entrepreneur, in terms of make sure you stay alive because if you don’t, then you cannot take care of the rest of them.
Advice for women founders?
Michael: I have a related question for you: we need more women founder role models for the next generation of open-source founders. Do you have any advice for women founders?
Amandine: It’s always hard for me to talk about this, because I’ve never felt that I’ve been hindered as a woman. I have the feeling that the problem is more that, early on, it’s like the work needs to be done at school level. It feels like it’s education, from young kids to understand that no, tech is not just for boys, and yes, everyone can be involved. And as a woman founder, I think we always tend to – it’s the things we hear quite a lot – but tend to not push and not take the seat at the table. I think we need to be careful about that as women. We tend to just like intervene when we think it’s necessary, but we need to be a bit more outspoken and push it a bit more. Because, otherwise, people around us who are just more pushy will just walk over basically, even if they’re less competent.
Michael: Amandine, thank you so much for joining me today. I know you have a busy schedule, probably flights to catch, so thank you so much for sharing your experiences on the podcast.
Amandine: Thank you for having me.
Michael: Special thanks to Amanda Brock and the whole OpenUK team for working so hard to launch the State of Open Conference. It was really amazing – thank you. Cool graphics from Kemal Bhattacharjee, music from Broke For Free, Chris Zabriskie and Lee Rosovere.
Next week, as I promised, Liz Rice from Isovalent, who’s going to tell us about the Cilium project and all the cool things they’re doing.
Until next time, this is Mike Schwartz, and thanks for listening to Open Source Underdogs.