SaaS Scaling Secrets
The SaaS Scaling Secrets podcast reveals the strategies and insights behind scaling B2B SaaS companies to new heights. Dan Balcauski, founder of Product Tranquility, leads conversations with successful SaaS CEOs, exploring their challenges, triumphs, and the secrets that propelled their businesses to the next level.
SaaS Scaling Secrets
Outwit your Opponents with Open Playbooks with Ian Tien, CEO of Mattermost
Dan Balcauski sits down with Ian Tien, the CEO of Mattermost, to discuss the fascinating world of messaging systems, open-source software, and the growing influence of AI in collaboration platforms. Ian shares his insights on the importance of being intentional when entering new markets, the exciting developments happening in AI, and how Mattermost is revolutionizing collaboration with its open-source approach. Learn from Ian's experience in building a successful messaging system that puts control in the hands of the users.
Guest Bio:
Ian Tien is the CEO and co-founder of Mattermost, a secure collaboration hub for technical operations. Mattermost serves tech giants, banks, governments, and some of the most vital and mission-critical organizations in the world with both commercial and open-source offerings.
He was previously an engineering leader in Microsoft Office, where he earned over a dozen patents. Ian is an alumnus of Waterloo, Cornell, and Stanford GSB, where he served as a teaching assistant for Andy Grove and Myron Scholes.
Guest Links:
Mattermost.com
Ian Tien on LinkedIn
Ian Tien on Twitter
Hello and welcome to SaaS Scaling Secrets, the podcast that dives into the trenches, the leaders of the best scale up, B2B SaaS companies. I'm your host, Dan Balcauski, founder of Product Tranquility. Today I have the privilege of interviewing Ian Tien. Ian is the c e o and co-founder of Mattermost, a secure collaboration hub for technical operations. Mattermost serves tech giants, banks, governments, and some of the most vital and mission critical organizations in the world with both commercial and open source offerings. He was previously an engineering leader at Microsoft Office where he earned over a dozen patents. Ian is an alumnus of Waterloo Cornell, and Stanford GSB, where he served as teaching assistant for Andy Grove and Myron Scholes. Join me as we explore Ian's stories, uncover the secrets to his success, and reveal the strategies he's employed to scale matter most. Let's dive in. Welcome in to SaaS Scaling Secrets.
Ian Tien:Thank you, Dan. Love to be here.
Dan Balcauski:I'm very excited for our conversation today. For those folks in our audience who aren't already intimately familiar with you, could you just briefly introduce yourself and tell us a little about your journey in the SaaS world?
Ian Tien:Sure. My name's Ian Tien. I kind of started out at Microsoft and Microsoft office. Really building sort of collaboration for enterprise, a number of the Microsoft products. Moved over to the strategy side, outlook.com. OneDrive were sort of, in my portfolio in the early days. They were called SkyDrive and Hotmail at the time. And then, did a transition, went to B-School and shifted gears. I started a video game company that eventually became the company now Mattermost. Where we went from online games to secure collaboration. And the funny thing is you look at like Discord or Slack, we all kind of started out with these video game aspirations and ended up in this. Sort of a real-time messaging space. Glad to be with an interesting cohort of other founders kind of figuring out in collaboration.
Dan Balcauski:All of us have those definitive moments that change our life. What was your superhero transformation moment? What's the, I'm Peter Parker, living my normal life a bit by a radioactive spider. Go to sleep, wake up and everything is different moment for you.
Ian Tien:I'm kinda like Peter Parker without the superpowers. So just, doing the best I can. I think it's, one of the themes that has kind of been throughout my career is the concept of community. I don't know if people realize how large the Microsoft community is, if they understand the Mattermost, the Microsoft Most Valuable Professionals program, how they get all these partners in Microsoft Office. To and Microsoft platform to work like 10,000 partners in the Microsoft ecosystem. It's a massive community. And, when I moved to video games, you know, doing that startup that became Mattermost, communities was so big in the gaming space. Discord is one of those, real game changers around how communities connect there. And in Mattermost, which really became from an open source community and moved to public sector and technical communities it's all about these ecosystems. It's all about, why do I, why do I. Join, why do I contribute? Why do I feel like I belong to something? And I think that frame has really been a thread across, everything that I've done professionally.
Dan Balcauski:You did join a very. Interesting community. At one point in your career, I couldn't help but notice that in your bio you interned at Trilogy Software here in Austin, Texas, which has had a lasting impact on Austin's tech scene and has some crazy alumni and stories that came out of there in the height of the.com bubble. I was wondering if you could share your thoughts about that unique period in technology time and what was that like here at Austin, Texas in 2001 at that company?
Ian Tien:It was an amazing time, like Joe Lamont and his vision for what Trilogy was and and the culture that he built, which was. It really unparalleled the time. And the thesis in my understanding, and it was, long time ago, it's really about like, hey, this, the internet's gonna change everything and all these startups are going i p o and there's all this attention. But, enterprises really have an opportunity for internet as well. And who are the sort of service delivery partners that can help with those transformations. And Trilogy was kind of built on that. And they would take the, I just remember how it was such a. It was such a hardcore culture there, right? You would, you have Trilogy University and this is all public information. You can, there's Harvard case studies on this, but people would be brilliant and they'd be really good. Not only technically, but also just socially just real great leadership, real good social skills, really strong technology. Just an amazing collection of fantastic people and what you've seen from that. Ground zero of Austin, Texas is, all these leaders. I can just kind of name the VCs and executives and, operators just coming from that trilogy culture. So, definitely I lived in Austin. I loved it. It was such an amazing experience. It's such a point in time and it's so great to see that that community continue to grow and and have impact around the world.
Dan Balcauski:For those of you who do not live in Austin, Texas, like I do, who may not be aware, so if you are in the Austin tech scene, you'll often hear reference to the Trilogy Mafia, much like the PayPal Mafia, Peter Thiel and Reid Hoffman and the rest of those folks. But Trilogy had such an outside impact in bringing top tier talent like Ian and others to Austin and really seeded a lot of Austin's future tech success. Austin maybe looks like somewhat of an overnight success if you think about the pandemic and, Elon moving Tesla here and all of his companies and. Joe Rogan, et cetera. It was really kind of trilogy 22 years ago or even beyond that, that really helped set a lot of that in motion. How did you go, so you mentioned you, you went to G S B and then started a v video game company. How did that happen into Mattermost? How do you go from video games into enterprise collaboration software?
Ian Tien:It's so I think what I loved about video games was, it was, there was a lot of passion and art and design and economics and, game balance and all these all these things you have to build to have even just an independent game, an indie game. We were doing massively multiplayer real-time strategy, M O R T S, and it was this really heavy system. And what we would do is we'd operate it all remotely. Using another product at the time, sort of before Slack, before Mattermost, that it was a startup and it was, real time messaging and file sharing and all the things that are in that category. And it got bought by a big company. And, after that acquisition, the quality really suffered and we would, it would crash. We would lose data. It had so many bugs and we were so frustrated. And when we tried to export out of it, we had, 26 gigs of our analytics and our data and our pipeline. They wouldn't let us export. And when we stopped paying our subscription, they paywalled us from our own information. So we were just, really trapped and really frustrated with a platform that was no longer working for us. And being engineers, we decided to kind of build our own, we had over 10 million hours of messaging in our video game platform. And we're like, Hey, we could build this. And as like an engineer, you write it three times. And the third, version of it was actually. Working and we decided to open source it. And what happened over time is it got noticed we had a very large Fortune 500 enterprise become our first customer. And and then another one from then another one from Asia. Then we got into our first government contract, and this was all, in, in probably nine months. And it was really clear we were onto something and we just decided, Hey, loved video games. It's kind of got a ceiling on how much an indie game company could grow for us, and this looked like a much bigger opportunity. So we just became Mattermost as we saw the demand for what we built.
Dan Balcauski:So you were dogfooding based upon the failures of a prior sort of workflow tool. You built your own. System to do that and eventually became enough for you guys just to wrap a company around. That's fascinating. For those of you who in our audience who maybe are not intimately familiar with Mattermost,'cause you just described it as a bit as Slack, but as I understand it, maybe it's not quite a direct comparison. Can you describe at its core kind of what Mattermost is before we dive further into the conversation?
Ian Tien:Yeah, there's a category of sort of, ChatOps tools or collaboration tools, so real time messaging. One-to-one or group communications. And on top of that, add content sharing like files and previews and sort of plugins and having a lot of sort of formatting and markdown and basic collaboration features. And similar to other platforms, we've got, video, we've got, screen sharing, auto calling, all those things wrapped in where, we're gonna be a little different is we really are focused on solving collaboration issues in complex environments and in complex environments. You really need a hub to bring everything together. Your legacy systems that have, just age old security and requirements, your new cloud native applications, how do you do, whether if, how do you bring these different disciplines together? How do you integrate all these systems and how do you have the right access controls and security? Our segments are really about defense, government and complex enterprises mission critical teams. We're saying no, this is not just like what's for lunch in the cafeteria. This is about like, I need to operate in real time and I need to do that reliably, securely and quickly with a lot of focus on process and sort of out of band communications. It's a code base that is, is open source and as a platform that folks can analyze and extend. So there's no other system that is as transparent and as extendable and is vetted by the sort of, three letter agencies that, that we're in. And that's kind of the core of Mattermost. One other thing that, that is kind of our that it's also part of the core, which is what our customers ask for is this concept of collaborative playbooks. So, you're getting you're talking and you're having conversations, but you also have this playbook, which is kind of a collaborative checklist with automation. And whether you're doing, an operation like release management or you're doing, you're taking vehicles and they're being deployed, there's these structured checklists with all the different approvals and complex logic that you need. And then what we do is we standardize that and integrate that with your communication systems so you're not jumping through like 18. 18 different apps. It's all one app. It's integrated, it's automated, it's secured. And that sort of collaborative playbook with that secure collaboration hub is the core of what we do. So, sorry that was a long explanation, but
Dan Balcauski:No, that's helpful. It's helpful. There's a lot of ground there and I think it can be helpful context for the rest of our conversation. There is something that really jumps out at me though, if I look at your background, Microsoft, Stanford Business School, and then I see open source. One of those three items don't seem to fit together. I don't maybe Stanford's teaching a different class of how to make money in open source. But it's very possible there. But Microsoft is definitely not known for that in a historical perspective. In fact, they were, very anti the Linux movement. So how. How did you end up deciding that open source was the right move? Especially as you're looking at jumping outta this, video game business, it sounded like it was somewhat successful. It maybe wasn't the path for you, but it was making money to an open source tool where out the fortunes of those companies are mixed at best, I mean, obviously are very happy about the open source movement in general. I think it's been a boon to a bunch of industries in, in many ways. But how did you make that decision?
Ian Tien:It's a great, it's a great question. Maybe just in two parts, kind of right now and then before. Right now, I would say that Microsoft is doing very well in open source. You know the GitHub acquisition, the GitHub folks are amazing, really like them. And they're, they've got, the vision for copilot and all the things that are happening and the contributions they're making to the open source community in, in terms of, their technical investments are actually pretty material. And what you find right now is they've really embraced, and then that's part of their. Sort of d n a is just embracing the, these platforms. Azure right now runs more Linux than it does Windows server by far. And you've seen this this movement as they've adopted cloud and this Azure and these sort of open and the GitHub acquisition has really changed a lot of the D N A at Microsoft. So I'd say one that's, that, that's certainly where we are today and. There's this kind of like the mirror image of that might be in the in the early days of Microsoft, people sometimes refer to that, the Steve Balmer era where, there's these quotes like, open source is a virus and G P L is terrible and all these different things. And where Mattermost came from, where our strategy came from was really the observation of how big a deal or how big a threat open source was to Microsoft in the early days. Open office was devastating, right? Like to, to Microsoft office. And if you look at the file format changes and you look at like how much antitrust shaped windows and office divisions, it was a really big deal and it really it really transformed the culture over there. And then sitting there myself as kind of a junior little person, watching this happen. It was really making me realize how powerful a force open source was. If it was going to take on, Microsoft itself, like this giant software company that shaped the world and worried about, wait, who, like these people working in their basements, like putting software together in their free time. Was this threat to, to one of the most. Material, most important software companies in the world. So what is that and what it is the concept of self sovereignty, right? Is that, so this concept that a country or a corporation or a person should be able to have sort of sovereignty on their. On their tech stack, on their technology supply chain, right? And they have, they should have the ability to read the source code and understand what's in there. And they should have the option of running and running their own technology stack and being sovereign versus, being sort of being just a customer to someone who can decide, what happens to their infrastructure they depend on. And it was such a powerful idea of the self sovereignty of open source, of this, idea that, hey we wanna have, we want a choice and we have wanna have options. And that's really unlocked. And then just understanding the ecosystem of, yeah, there are people who want stuff out of the, like, Microsoft's a, it's a wonderful, platform. It comes outta the box. It's priced really well. You have all these sort of fun these, pieces of functionality you get. And the difference with open source is, you get, it takes a little bit longer in open source to, to build all those pieces, but once you get it it kind of becomes pervasive and you can kind of customize it the way that you want. And you have it in a sovereign, in, in a sovereign format. So it started with operating systems like Linux. You got databases like Postgres and MySQL. It went up to, you've got Kubernetes as the virtualization layer. And right now we're going to application layers. You've got companies like GitLab that, they've got the whole application suite of DevSecOps. You've got Mattermost, which is secure collaboration. And once you've got these open source offerings in that layer you really change history really. We say that no one's really gonna start competing against open source Linux for that layer. No one's really gonna compete against Kubernetes for that layer. And same with the applications that are coming up. So the Now Microsoft is certainly embracing open source and the before time, what we noticed was, just how powerful open source self sovereignty and community can be with a seat at the table at the global sort of software infrastructure,
Dan Balcauski:That's interesting
Ian Tien:the global software instructure, at the global software infrastructure table.
Dan Balcauski:Absolutely. Well, And Satya he'll go down in the record books as one of the best CEOs of all time. He has definitely turned the ship there and I think made them more open source friendly than yeah, they ever were in the Steve Balmer bill Gates era for absolute sure. So all props to him and what he's. Doing right now is just unbelievable putting Google back on their heels, I think for the first time in in a long time. So it's exciting to watch those two giants play it out. There, what you talked about this self sovereignty idea was this principle concept that was driven, sort of by you and your founders fundamentally, that you were dedicated to that? Or was that something that learned as you started looking at, who was adopting this and where you thought, there was a market opportunity? Was it a bit of both? Like how did you, how did that come about to be such a found foundational principle of what you're wrapping matter most around?
Ian Tien:It was totally by accident, right? Like you just, you don't know what it's like. It's like, having our platform, our own collaboration platform, which we got from a SaaS startup, just start dying on us after, just start dying on us. And just having no ability to export our data. And no ability to actually do our work was really frightening to have the opposite of self sovereignty. So built it for ourselves. We it's if you kind of hurt your arm, you kind of gonna baby it. And we kind of built up this massive muscle of like, well we kind of overbuilt this open source, collaboration platform. And then enterprise just, really wanted it. And we just saw the opportunity. They wanted the same thing we wanted, they want independence, they wanted transparency, they wanted control. And they wanted something that could adapt to their needs. So really came from just, it just came from a bad experience, right? Like when you're on the other end of it, you're like, this sucks. And then really just solving our own problem same challenges.
Dan Balcauski:that makes total sense. No, I appreciate that. And it was sort of that, that Robust emotional experience that sort of lives with you. It's like, never again will we do this trust our data to somebody else? It won't get locked out of it when we need it back. Look, I you've touched on a bunch of threads and I think one of them is really interesting is, how do you sort of grow in open source business? And I think about these ideas of growth levers, really, there's only way three ways to grow a B2B SaaS company, which is, acquisition, retention, and monetization. As you guys sort of went into this world, you've. You go into, you have this open source platform, it seems to be getting some traction, how do you think about monetization of open source? How'd you guys start? Are there things that you've learned along the way that apply in this domain?
Ian Tien:Yeah, it's really, it's really just under, like, when we got an open source, like, like we'd never done it before. Me and my co-founder were ex Microsoft from the Balmer days, and we didn't know the first thing about what open source license to use, what did it mean? And we found this wonderful, lawyer and I'm like, help me understand it. And she's like, I like it's gonna be too expensive for me to, she gave me a book on software licensing and I just used that to start figuring it out. And
Dan Balcauski:You could have used chatGPT back then. Now you'd be like, Hey, could you just summarize this for me? Tell me what I need to know.
Ian Tien:Yeah. Yeah. So, so yes. It takes a little bit of time to educate yourself in open source. There's only three ways to make money. There's sort of three main ways to make money on open source. And two of them are not very good. One is services, right? Support and services. Now that's very classic. That's sort of open source 1.0. And so that's one. The second and the first one's kinda like, it's kinda like red hat, right? There's a lot, there's support contracts plus there's, IP indemnity. But there's, that's kind of 1.0, it's not, I guess you could scale it for certain businesses, but it's a tough business because you're actually incented to have your system maybe not work perfectly, right? Because otherwise your support revenue goes away. So that's one. The second is cloud hosting. So you look at like, WordPress and like the, that's a, you can, anyone can run it, anyone can run WordPress, but wordpress.com, runs it better. And, this is from the creators of WordPress that's a second model also. Undifferentiated open source being run better by the people who made it, again, creates these kind of weird incentives and it's, and margins are kind of like maybe a little tricky. The third version, which is kinda the most popular version these days for venture-backed companies is open core, which is that you've got an open source solution for one use case, and you let that just kind of dominate that use case. And then you've got a commercial version for a different use case. It'll be more, it'll be richer, it'll be for a different buyer audience, ideally. And and that's how you do it. And that's a, that's the method we used at Mattermost, which is, hey, there's an open source version, it's called Ma Mattermost Team Edition. You use it for, small teams and it does it, you do collaboration and you integrate with all your things, and you don't have to pay anything for it, and you just run it when it gets to a certain scale. The developers will run Mattermost. But then the IT manager's like, oh, wait a minute. I've got like 18 teams running Mattermost, and they now want me to manage it for them and for me as an IT manager to run it while I need single sign on and active directory and user sync and metadata. I need these compliance policies. And then they go to our website and they're like, oh, for all the things that you want, you can buy this, subscription. And then that's how they flip over to a to paid version. So it's really let that open source version, adopt and deliver a lot of value. And then for a different persona, let that person, solve a new class of problems in a paid package.
Dan Balcauski:That makes sense. So you've got kind of the core well just let developers use it because developers are gonna use something that they, that's open source, so we'll just let them do that. And then have these other port paid add-ons for or paid tiers for other buyer personas. I guess, it's clean how you lay it out. Now, I'm, I'm wondering, if we go back four or five years ago as you're making this transition for video games to not, was that path clear? What'd you learn along the way? Was it just like, oh, of course. It, managers are gonna want, once the amount of most team editions proliferate in organizations they're gonna wanna pay. How did it evolve over time? What were the lessons that you learned along the way?
Ian Tien:I think I think just finding models. So I think, Mattermost is actually bundled inside of GitLab, so if you use GitLab Omnibus, you actually have the free version of Mattermost. And, we've worked with GitLab for many years since the start of Mattermost, and they basically said like, yeah, just try, just use our model. It's like, it's fine, right? It works. And that's the kind of the ethos of the open source community. I talked a lot of open source founders, know, so we're backed by yc, I mean, talked to a lot of yc open source, open core founders, and they asked the same questions and they gave the same advice GitLab gave us, just copy everything that. Look at all of our stuff. It's all open source. Use it, use what you like, and if you don't like it, change it. But it's all there. And that is the, that's sort of the spirit of open source, right? It's like, hey just go use things that work and adapt things that are different and it's your choice. Whether not you keep those proprietary or you share it back with the community. So those are all sort of your choices. And I think that takes a little bit for folks to understand that don't come to open source. There's a lot of ego. They're well, I can't do it the way that they do it. And we're like, no just run Linux. Don't get creative. Run Linux, run Postgres. Run K eight. Don't try to like, like people wanna. And for the business model, it's the same thing, right? Just like you're open core, look at those open core licenses, just copy the ones that we want. I was on a call like last week, same thing. They're Hey, how do we do this? And it's like, here, just read it. I love the open source that way. It's a great, it's community. It's time saving, it's efficient. But what you can't do is you can't come with ego. You can't come with like, well, not built here. I wanna do, I wanna be special. How dare we copy that other thing. This is, it's about build on other people's successes. Build on the comm, success of the community.
Dan Balcauski:Yeah, Well, you mentioned GitLab too. I am always impressed by how open they are with, as a pricing professional, how open they are with their entire pricing philosophy. If you go to gitlab.com their pricing page, they've got a whole manifesto, ethos, I don't know what you would call it that describes all the trade-offs, the decisions they make and how they go about it. And so you're absolutely right. You don't even need to have a conversation with some of these founders. You can just, they just publish it on the open web. That's how open it is. I'm curious, you mentioned that a lot of your customers they're really data conscious, right? Self sovereignty and the three letter agencies that you mentioned, I imagine that some of these are quite, maybe they want their instances, in air gap networks or et cetera. I, as it pertains to pricing monetization. Does that create any challenges you guys have had to surmount with sort of monitoring, license enforcement et cetera? Any lessons that you've learned in that domain?
Ian Tien:Yeah, certainly air gap networks and, not having telemetry'cause like it's not possible is. Yeah, I think it's par for the course, right? Anyone who sells an air gap networks has had the same problem. And what we do is we have what's called an open license, which means you don't have to come back to us, to buy more licenses, to unlock more seats. What an open license does is saying, Hey, we trust you, right? Here's the terms. And depending on like the size of deployment, there'll be a period that we just true forward. So it's like, Hey, you start with 10 users. Maybe you want to add five and you don't have to come back to us to buy a license to unlock it like a SaaS product. You can just deploy them and then at the right TrueUp period you'd be like, Hey, I deployed five more. I'm gonna true forward. So we prorate that for the rest of the year, and then you just pay that prorated amount and you renew for that amount. It's, GitLab actually does the same thing. They've changed theirs a little bit over time. But if you look at Mattermost and GitLab's licenses, it's pretty much the same. And it's built for both self-managed and cloud managed networks.
Dan Balcauski:There's a lot of trust there. I imagine there's maybe some CFOs who get a little nervous with that level of non oversight. But but no, I mean, yeah, there's some of those situations you're like, well, if you wanna be in that business, those are some of the terms. The other thing it was interesting was so on, there's monetization and there's also acquisition. It sounds like, you guys had talked about, the ability to extend the platform, and I'm sure there's, many other systems that, share. Data with, or have to be interconnected with, how have, how has that influenced your thinking around Mattermost, go to market motion, do you go to, go direct for most of your business or through partners? Like, how have, how has you think about that? How has that evolved over time?
Ian Tien:Yeah, we're definitely a majority channel driven business. The reason is we're, and matter most, the open source version is in 20 different languages, right? It's it's all localized and it's used around the world. And what happens is folks will need that, that they will need to buy that subscription to the paid version. But we might not speak, Japanese or Korean. I mean, and now we do, but like in the early days we didn't. So we always had a reseller network that would be able to, help us, trend, not only transact, but also support in, in, in different countries, in different languages, in different markets. And that's actually preferred for a lot of enterprises. Ms. Enterprises have big MSAs and they have all these agreements. The right channel partner. We get this question asked a lot, which is like, well, how do you deal with these like giant MSAs and negotiate and like the price point has to be so big, you need to champion all that. And when you actually go through some of these channel partners they can make it super easy'cause they just they do that for the enterprise. They are the a p i between sort of startup and big enterprise when they wanna procure and they'll help take care of some of the terms in the middle. And then sort of everyone wins. So I think with a open core, bottoms up strategy, so, you try it open source, you get the value delivered, you need a little extra value from these advanced features, and you're willing to pay for it. And then you make the transaction easy because that enterprise can go internally to their channel partner sort of like they're sort of like a p i into all these businesses and they can procure. And do it in a standardized way. And then we support and we support that we can support a lot more customers through channel and we can be a lot more easy to access.
Dan Balcauski:It boggles my mind because, you had this I guess you're in the video game space, obviously, Microsoft has a big partner network, et cetera. But, was this something that as you guys were building matter most that, I talked to so many SaaS leaders who just. Partnerships and going through channel is just kind of this foreign concept. And so it's a really interesting conversation for me because it, it seems like you guys are so much more advanced than, it's like I, I see people trying like, oh, we're gonna dip our toe in a channel program. And it's, half-baked and that they have all sorts of problems'cause it is half baked. Is this something that, like how did you find these partners that were the easy button? Was this from experience, from connections that you had, previously or like how did Did you bring in an owner executive who was, who, was an expert in building these enterprise partnerships?
Ian Tien:No, we just did it ourselves and and how kind of, we brought some advisors in who, who had done it before. So we brought advisors, but then we did the work ourselves and we're a little bit lucky coming from Steve Balmer era Microsoft, which largely on-prem. 96% of Microsoft revenue came from partners came Microsoft doesn't deliver Windows. They don't, you don't go to Microsoft and ask for a quote most of the time, and unless you're like big enterprise. But all the SMB and mid-market would go through distribution partners and channel partners who have services also implementing Microsoft and you know, windows and Active directory would've Windows certified professionals. So 96% of revenue at the time of Microsoft was going through partners and that community. So we just like, like that's the way that we knew how to sell software. So we did take some deals on direct, but the channel based deals, value added resellers versus system integration partners. The the margin you give is gonna be a little different based on are they doing, local support, eight to five local time zone. There's these, just these templates for how you do. Channel relationships, they're going to differ by country and by region and by sort of service level. There can be distributors or full systems implementers, and you just need to, it isn't complicated. You just need to fit into those patterns.
Dan Balcauski:Yeah. I think the thing that trips up a lot of SaaS folks when they first get into this right, is different than the perpetual world. Perpetual world. It. At the end of the day, you may not know who your end customer is, in the SaaS world that's, it's kind of been baked into the subscription model. And then, working with a partner, I think there's often questions of who owns the relationship, do we know who the end user is? I guess how have you guys, how have you navigated that question? Has there been I guess is it a matter of hey, there are partners that really understand the subscription model and know how to work with. These newer businesses versus the, I guess the Steve Balmer era maybe resellers?
Ian Tien:I mean those, there there's certainly challenges. Challenges and certain things you have to do. Make sure if anyone listening you're doing export, make sure you have good export compliance partner that is gonna make your that's the one thing everybody super careful about in terms of who your partner sells to and how they sell it and what you can export outside the if you're in the US based company outside the us, make sure you get export compliance. Correct. I'd say in the partner models just work with channel partners that are like, Are really good. Don't work with like sketchy ones. So work with the best when you can. So easy one for us is GitLab we're distributed through GitLab, right? Like you can get an open source version. So we just, started with their channel partners, right? It's easy'cause then it's great for them'cause they can co-sell those two things together. That was an easy win. And then we kind of, blazed our path in some areas in defense and federal government. We're pretty, we were very lucky to work with Carahsoft, which is the number one, distributor into federal. They're an amazing partner. Very high quality. Just awesome. And, they're helping us distribute and And fulfill and market, our services. The one, so there's two parts. I mean, there's many parts, but there's two big parts of these sort of partnerships. And channel one is the inbound fulfillment, right? So we're product led growth. And very often people will come to these partners to buy. That's the inbound motion. And our direct sales team does work with those opportunities to make sure everything is flowing the right way, that we're understanding their needs and we're supporting these larger deals. And they get fulfilled by channel all the time. Or channel actually, webinars and things like that. They'll kind of bring in. Then the other one is sort of like the outbound, more outbound, version. And and there's, and there's a sort of the middle where you're doing demand gen and things flow in. We're probably more on that inbound fulfillment. We do a number of that, sort of like that demand gen, drive awareness and bring folks in. We do not, we're not really yet at, and we wanna get there. And that solution sell with a partner, with a full stack partner solution. What we're not in that sale deliberately. What actually happens is a lot of time these large systems integrators will include Mattermost as a component. To a bigger solution and they will solution sell with us kind of in the bill of materials, but we don't. But we wanna make that more deliberate over time.
Dan Balcauski:No, I just there's so many. I could just imagine a matrix of different options that you have to think through and I think this is one of the reasons why, you know if you're not sort of intentional like you guys have been, in finding the right partners really understanding your go-to market model, what makes sense, Hey, are we gonna have g do we need help because of Knowledge of local markets and, partners get, access to that market versus, if there's, is a lead comes inbound and the partners worked with that customer before, right? So there's all these things I think just people don't think through until they're in the thick of it. So I really appreciate you sharing your experience there. I want to quickly pivot. I know that you recently had some pretty significant announcements around dipping your toe in the water around ai and it's one of my absolute favorite topics. I am curious because, You've talked about, all of the, know, da you know, self sovereignty and process. And as I think about what's going on in ai, it's like, okay, that's a big problem.'cause there's, you ask LLM, you know, to give you an answer and might, depends on what day of the week, depends upon some random number generator, what what answer you might get from a machine. So I'm curious how you're seeing. Your customers or your market adopting ai and how is that influencing how you're thinking about, Mattermost and your roadmap?
Ian Tien:Great question. I mean, there's been a lot of there's been a lot of demand for sort of a AI exploration, AI solutions. There's probably not a, an engineering leader that hasn't, hasn't made it a priority to take a look and at least investigate. Very often it's gonna be sort of smaller teams smaller teams of a larger organization doing that initial work. And, the way we think about it is that there's gonna be, different segments and layers. There's gonna be sort of, general generative ai, which is hey, I've got LLMs. I can ask basic questions, I can do summarization. There's a set of sort of general operations and that's, that's getting. Built in Mattermost through a platform we call open ops. Some thread summarization, meeting summarization action items being able to draft responses, being able to, translate, all those things are available. In, in the open ops sort of platform. And that's kind of like the basic layer. The sort of next layer beyond that is, is what we think about as sort of domain specific, right? In our defense industry, a lot of folks are looking at, for example, acquisitions. Right? The acquisitions workflow is kind of a good use case for Mattermost. You're like, Hey, There's all these regulations I need to follow. There's all these steps in a collaborative, playbook that I need to go through. And I really need to go and scan and analyze a lot of documents, right? So it's like, okay, there's this, 20, there's 80, 25 per page repo proposals. There's a saying in federal government reports aren't read, they're scored. So you have the analytical, you have a framework you have to go through, and it, the, there's a lot of, there's a lot of efficiency that can be generated through these systems. Be domain specific. And there's, there's fine tuning. There's, All these techniques you can use to, know, openAI just just announced fine tuning on a sort of per instance level. So there's all these things you can do on the domain specific level beyond general. The third piece is really sort of what we think of as sort of line of business, which is, you're gonna be not only sort of analyzing but sort of fully operating and you're gonna have these mis recommendations, but you're also gonna have sort of agency. You're gonna be able to have these conclusions that these, that the machines are drawing and actually put those into action and replace, some of the more More basic tasks. There's this person Cassie at at Google. She's like chief data scientist, and she invented this word called thunking, right? So there's thinking, and there's thunking. And thunking is all the stuff you need to do. Like, the data gathering, the pre-read, that there's kinda schlepping of information. That stuff you have to thunk before you can actually think. Her view is like all these funking activities will eventually be replaced so we can do the high quality important work.
Dan Balcauski:No, it's I definitely have been there. It's the 95% of a data project is just, yeah. Cleaning the data and getting it ready for the analysis you had in mind the whole time through. It makes it very difficult to estimate data projects. I see that doing it myself as well as a product leader. That's fascinating. So the open ops platform, Kind of your base level sort of, operations of, summarization, et cetera, what you might expect out of, what people are doing with G P T or any of these other tools, a baseline layer. And then sort of this, domain or department specific levels of capabilities that are More trained on kind of what is that specific domain and what is the data or required there? And how do those systems work at a more detailed level, that open ops platform that you specified, is that just an open source project in and of itself? Is that something that Matter most has developed? What how does that fit in?
Ian Tien:That's openops.Mattermost.com. You can go to it. And it's really just our approach at supporting these AI workflows. So Mattermost itself is not gonna have LLMs inside of Mattermost Like, we're not shipping that. We're not, that's not. Our space. We're not in that layer. What we are is we're the collaboration layer, and what we do is we support AI accelerated workflows. Customers, one customer the other day was describing it. They're like, well, we, this is our inner circle ai. Scenario, right? Like it's not in that they might be using Microsoft Teams or some of their general collab piece and they've got some AI that's always gonna be built in and they'll, they'll do certain things with it. But there's gonna be like inner circle data, the stuff that's really confidential and really compliance heavy, and you really want to have that in something. We have full control. So Mattermost is a collaboration layer. You've got full control, right? Every. Every bit or byte that goes in the database like, is yours, no one else can see it. You have complete privacy and for the folks that are using these self-managed models, whether it's a w s bedrock or it's LAMA two for meta, if you're looking at a self-hosted, self-managed model, well you kind of need a, the similar self-managed collaboration platform for that to plug in and have that all as sort of your inner circle workflow. And because. The data that come, the data you put in and that comes out is just as sensitive, right? And. The other piece that's opening up is like, you really do need personalization in these AI workflows. So Dan, like if we are taking this recording and we run it through, just, whisper in the standard scanners and it comes up with a five bullet point summary, your summary should look different than mine for it to have real utility. Right. And the only way we can do that is to have knowledge, private knowledge. Of what you're interested in what I'm interested in based on our behavior. So you can have this personal data store that tells the system what we're interested in. We gotta pass that back with our query to get the right summary that's gonna be the most utility to us. All that private data. For some enterprises really needs to be under a hundred percent under their control. And that's our market is that private ai, that inner circle workflow, that confidential sensitive data that really needs to be processed and and our workflow is accelerated.
Dan Balcauski:Ian, I absolutely love talking to you. This is your grasp of, both the business side and the technical side. I feel like I'm sitting here having a conversation at a, with a level of a chief engineer but also that we could flip over to the business side. And it's just such a rare combination. It's amazing. I could talk to you all day but I do want to wrap things up here with a couple of. Quick, rapid fire questions. Look, obviously you've been very successful in your career. Nobody who has achieved a level of success gets, to the mountaintop on their own. Has there been a close mentor or other leader who's really helped you on your journey?
Ian Tien:I'd say, I think you have to start off with the bio. I worked for Andy Grove and anyone who's worked for Andy. Grove will just, I think always just reference their time with him. And and the this, and the nice thing about that is he's got a lot of books, high output management, Only the paranoid survive that people can go to and kind of get that same sort of foundation. So if I could recommend to any sort of ceo or leader high output management would be would be my number, would be my number one. And Andy was for anyone who's ever worked with them. Some of that really echoes on in terms of their. What they gave to the world in, in, in terms of models and frameworks and how to think.
Dan Balcauski:What an amazing leader. I've got high output management over there on my bookshelf. So absolutely recommend. To anybody who hasn't listened or hasn't read it before or listened to it. But but an amazing book on management and very practical not fluffy business. It's like, here's how you lead an organization. Boom boom. Not that well, you get a normal business book. It was amazing that you're able to build that relationship with him. That's fantastic. Look If I had give you a billboard, you could put anything on any advice for other B2B SaaS CEOs who are trying to grow scale their companies. What would you put on that billboard?
Ian Tien:I think from the conversation today, I'd say, frame everything in community. I don't think people understand or appreciate that the power of a community is, whether it's community of your fellow CEOs or a community with your board members or your investors, a community with your customers, a community with your partners, a community with, your internal staff. So if you frame everything community, you've got this operating system for sort of, for organizations and for ecosystems. And, just having been through open source, having been through Microsoft, having been through sort of the online entertainment community, this is, that, that's what I would share.
Dan Balcauski:Frame everything in terms of community. I love it. Well, Ian, if folks wanna learn more about you or Mattermost are there places they can go online to, to learn more, follow your social media, handles, anything like that?
Ian Tien:Yeah, I'm on Ian Tien at On Twitter and LinkedIn. Just my full name and Mattermost.com is where you'll find, all our best stuff.
Dan Balcauski:Excellent. We will include those links in the show notes for listeners, everyone that wraps up this episode of Saas Scaling Secrets, a massive thank you to Ian for sharing his journey, insights, and invaluable tips for our listeners. If you found this conversation as enlightening as I did, remember to subscribe so you don't miss out on future episodes. Until next time, keep innovating, growing, and pushing the boundaries of what's possible.
