Building a Strong Open Source Community With Corey Hulen (CTO, Mattermost) - Percona Podcast 45

by Corey Hulen, Matt Yonkovit

Link to listen and subscribe: PodBean

Engaging users and contributors is key to building a strong and vibrant open source community. At Mattermost, the team spends a lot of time ensuring their community can be successful by making the contribution process as easy as possible and providing a welcoming environment. Corey Hulen the CTO and Co-Founder of Mattermost joins the HOSS to talk about his journey to the open source space and how he has been helping to build a culture internally that promotes and embraces the community as key members of the engineering team. Listen in and get tips on how to grow your contributor communities as well as learn about the cool things happening at Mattermost. Whether you are in your company’s OSPO (Open source Program Office), DevRel team, or an open source maintainer this is a great talk to learn about some cool ideas and best practices.

YouTube

Link: https://youtu.be/5sYSKxF97OM

Corey Hulen

CTO at Mattermost

Corey is CTO and Co-Founder of Mattermost, Inc. which is an enterprise messaging workspace for teams to collaborate securely and effectively. He previously founded Tempo AI, an artificial intelligence startup. Earlier in his career, Corey served as engineering manager and architect for Microsoft Office. He was also an architect for VerticalNet, a publicly traded enterprise software company providing supply chain management solutions for the Global 2000.

See all talks by Corey Hulen »

Matt Yonkovit

The HOSS, Percona

Matt is currently working as the Head of Open Source Strategy (HOSS) for Percona, a leader in open source database software and services. He has over 15 years of experience in the open source industry including over 10 years of executive-level experience leading open source teams. Matt’s experience merges the technical and business aspects of the open source database experience with both a passion for hands on development and management and the leadership of building strong teams. During his time he has created or managed business units responsible for service delivery ( consulting, support, and managed services ), customer success, product management, marketing, and operations. He currently leads efforts around Percona’s OSPO, community, and developer relations efforts. He hosts the HOSS talks FOSS podcast, writes regularly, and shares his MySQL and PostgreSQL knowledge as often as possible.

See all talks by Matt Yonkovit »

Transcript

Matt Yonkovit:
Hi, everybody. Welcome to another HOSS Talks FOSS. I’m the HOSS, Matt Yonkovit. Today I’m here with Corey Hulen, who is the CTO of Mattermost. Hi, Corey. How are you?

Corey Hulen Hi! Thanks for having me. Glad to be here.

Matt Yonkovit: So Corey and I worked together briefly a few years ago, and we have kept in touch over open source conferences, and knowing that Mattermost is so dedicated to the open source space, I thought he would be a great resource to come on and talk a little bit about how they’re changing some of how people operate DevOps and how people operate their systems and talk a little bit about his experience in the open source space. But your background isn’t in open source until Mattermost, right?

Corey Hulen:
Yeah, that’s true. I guess I wasn’t, at least I didn’t like working in it full time or getting paid to work in it and I was doing random small stuff here and there, but nothing, nothing, nothing to this scale.

Matt Yonkovit:
You worked for Microsoft before they loved open source, right?

Corey Hulen:
Yeah. Yeah, I worked for Microsoft before. They love open source. I’m coming from that environment and seeing them today and talking to some of my friends. They’re like, I’m truly impressed by like, I know a lot of people like, yeah, a lot of people consider them the evil empire. But I know what I’m and it was like, they’re so bad back then. But just knowing that sort of mentality they were back then compared to where they are now. I’m truly impressed. But yeah, so I did that really, way back in my career, and then some other startup stuff, and then you’re anonymous.

Matt Yonkovit:
So coming from, you said, the evil empire? I did not do a great job. But Corey said, the evil empire. So coming from the evil empire, into the open source space? You know, how did that look like that? That had to be kind of like this is weird. This is a little strange. What was it?

Corey Hulen:
Yeah, I think it definitely is. I mean, I think there’s a lot of similarities or things that Microsoft does, like building community and stuff like that, that they do really well. It’s not necessarily an open source community or competing open source or sort of source available stuff, but like they are, they always had that kind of motion, that they’re really good. I mean, they’re also good at the sell side of the two. But, I think that there are a lot of similarities there. But yeah the big piece missing is probably the mindset of the mantra, which I think has definitely changed about being open source and stuff like that, it was definitely more of a, even though they didn’t really do well, it’s a controlled relationship. So it’s really fun to see that experience and apply some of that experience to sort of outside. I think they’re really good at producing software and teams that produce software and how to innovate and stuff like that. And like I said, they have to do it at a scale that very few companies can do it at. But, uh, but yeah, I think they’ve been in price. I think their learning curve for the last several years has been around open source. Space has been pretty old Microsoft by now.

Matt Yonkovit:
Okay. But from a user perspective, right, so let’s talk about you like, did you have an aha moment where you’re like, What the heck is this? This is crazy. Like like, this is not how I’ve, I’ve done things in the past.

Corey Hulen:
I don’t. Yes, but honestly, it wasn’t until I left Microsoft, for a lot of reasons back then that a lot of it had to be built here because of liabilities and legal things and all that stuff. Right. So it was very much we weren’t allowed to even look at external source repositories, because they’re so scared of being sued and whatnot. And so, for me, that was a big aha moment when I ended up leaving, just having that freedom to like, oh, okay, like, look at this, there’s all this other stuff out there, that’s great that I can borrow to learn from however you want to describe it. And I think that’s a mindset that has radically changed, like, like, this is a few years ago now. But I had a friend there who showed up, we were meeting and he showed up with a MacBook. And I’m like, That is like the weirdest thing in the world. To see a Mac or a Microsoft employee show up with a company laptop like a MacBook. So I think as I said, that’s not to say open source with it. But I think this mindset changes in terms of something.

Matt Yonkovit:
So when you started the first open source project, which was Mattermost. And before Mattermost, it was not Mattermost it was the gaming company so as you started to start to develop your like that first iteration of your open source project. How did that go? Like, what did that look like? Like? Did it feel weird? Yeah,

Corey Hulen:
I mean, it definitely felt really weird, especially for me, like, Come like, like I said, I’ve done some small contribution stuff here and there, nothing to brag about, or nothing major at all, and to be on the other side of that offense in terms of going through this process of, hey, we’re gonna open source all this stuff. And we did it for some strategic reasons, right. One of them was, we just thought like, when you think of Mattermost Back then we thought of it as a platform. And if we wanted that platform to be ubiquitous, like we thought open source was the way to do it, right? Just a ubiquitous, easy-to-use platform that I won’t get involved in. So that’s one side of the reason, like sort of the logic behind doing it. But then there’s the act of doing it. And then there’s the not only act of doing it, but going from one of those developers who always contributed to private repositories cared more about probably the outcome versus the actual like little quality of the codebase, sometimes, right? You end up having, at least for me, a real Tiffany moment of like, oh, wow, like, we’re gonna open sourced this stuff. And like, it has my name written. I better go through and clean it up. So and so that was a really interesting experience for me, because like I said, I’ve never done it on that scale. And I think the more interesting experience for me, and this was the aha moment, was when we, when we did open source, just saw so many people wanting to come to contribute and help and look for the source code. That was an amazing thing to me, in the sense of, like, I was kind of like, almost, I mean, inside in hindsight is like, the core, it’s obvious, but like, at the time, like this is I don’t know, six years ago, now, maybe seven years, maybe got longer and more than five for sure. You know, at the time, and like, I just didn’t think people would be interested, right? It’s just, it’s like, okay, whatever, go start it’ll be there. If people can look at it, blah, blah, blah, but I didn’t think it’d be that big of a response. Right? And that it’s the same thing. It even causes more fear and anxiety, like, okay, there’s a lot more people looking at this, that I would have thought like, let me go clean some more of it up. So there’s a lot of that kind of cyclical process going.

Matt Yonkovit:
So Mattermost has a pretty active contributor base, which is kind of cool. Right? You know, not all projects have that like so quite honestly, I’ve talked to more projects that are desperate for contributors, then that are like, Oh, this is great. I have so many I didn’t expect this. Right. It’s the opposite of what a lot of people experience.

Corey Hulen:
Yeah. I think in the early days, it was probably more sort of not. I think it’s kind of two things. I think it was the space, right? People are always interested in communication platforms and stuff like that. That’s one. I think it was the technology. I’ll be honest, like, when you look back today, it’s really the easiest choice. But back then we made these choices being a go-based server on the backend react on the front end like those were very new and novel choices. Five, six years ago, yeah, go has been around for several years, but I wouldn’t call it it’s nothing like it is today. It was definitely not mainstream when it made some of those decisions.

Matt Yonkovit:
So good opportunity to learn that.

Yeah. Right, so copy all of our bad mistakes, right? I think the same thing, like React Native, is another example. We’re one of the early adopters of that, as well. And I think Facebook uses us as an example at their f8 conference because I think we’re one of the largest open source at least React Native projects, right. And so they have a bunch of closed source ones that they can’t showcase on stage or talk about or don’t really want to, but ours is open source so they can do a lot of that kind of stuff as well. So yeah, I think it started off like getting lucky having the right alignment. But then we’ve also poured a ton of energy into that, like, we do a lot of what we like to call community buddies, nurturing and mentoring that and then honestly, the biggest thing is, is directing our audience. So we do it through how we want tickets, we’re really good about trying to create Help wanted tickets and get them published out there. So people can discover easy ways to contribute. And we even ranked them or labelled them, like easy first contribution tickets, stuff like that. I think those days that people don’t realize, like, that it takes a lot of work. And it’s a lot of energy to do those things. And it’s just constant energy. It’s just sort of like a hum that you just have to keep doing. And I think that’s one of those things that people don’t realize, like, you have to keep up that constant energy to keep the community engaged. And we go through spells where we’d like more contributors and less. But, I think overall, we’ve been really lucky.

No, and I think that that’s one of the interesting things is especially from a community perspective, the consistency matters so much externally, it’s easy sometimes for you to dismiss it internally, right? Because you’re like, ah, yet another thing that I have to do maybe I could just let it sit for a week or two or whatever. But you don’t realize how many people externally are counting on something like that to happen on a regular basis or to be able to go out there.

Corey Hulen:
And I think being reactive, like another really great thing that I still probably still pride ourselves on, we typically get back to a community pull request. We always had a monitor we wanted to get back within one business day. I don’t know what our stats are today, but I guarantee you it’s close because even if it’s three, four days that’s like light years ahead of other projects where we have people who come to us and say oh yeah, I submitted this project and it took like three months before somebody reviewed it, right? We try really hard to be very quick on that that review cycle that review period and make because like I said those contributors there’s they’re choosing to spend their free time, their energy working on your project, like the least we could do is be receptive did that, like I said, on average, like our response time is really good. And that’s one of the things I’m constantly proud of because we hear that all the time when I randomly meet with people in our community and ask them, well tell me some good things about us. And they’re like, oh, like, it’s amazing like I submitted a pull request to you all, and it was reviewed and at least reviewed may not have emerged, like, I’ve gotten feedback about it within a couple of days. And then people submit to other projects, and it can take months or even years.

Matt Yonkovit:
But like, so this is an interesting thing because I’ve talked to a lot of people who have those backlogs of PRs that are 800 hours long. And I think that it’s an interesting debate, right? Because I hear this a lot. We’re trying to hire engineers, we need more engineers. So we’re focused on our core products. Right, and we’re focused on getting out the next release. And so we don’t have time for those PRs. So how do you find the time? How does the team find time for those? How did you bake that into the process early on?

Corey Hulen:
I think we, I mean, I think it’s just a mindset as a company, like, like like, the executives believe in, I believe, and it’s just one of those things we push, I’ll give you the best example like, and it’s like one of those things until people have epiphanies. And so we have, we kind of source recruits from two different areas, obviously, one is from our open source community, which we love, because they can already hit the ground running all that kind of stuff. And the other is just not open source people, right? Good people who come in through recommendations or referrals, whatever, right? And the mindset of those students, and there’s not one or another like, but the mindset is very different. Because if somebody comes from the community, they kind of understand, like, what needs to be helped. But the funniest thing is when I find someone who’s not from the community, and they start working on something or doing something, and I’m like, hey, no, be great. That’d be an awesome community campaign, right? Like, create a campaign around it, go create 5-10 15 Help wanted tickets. And let’s see what happens. And most, and it takes a lot of convincing sometimes, especially people who are not coming from the open source community, like, just, I just, it’s not like a requirement. It’s like, you must do this. I just keep having it like, Hey, have you tried to campaign around that yet? No, no, I just keep working on it and finally, somebody will buy it. And they’ll be like, Okay, we’ll create a campaign. Corey, stop bothering me. Here are 10 Help wanted tickets. And then it kind of sits there. And it’s kind of typically it’s like crickets, right people like, and I keep telling people, no, no, this is how open source works, you got to be kind of talking to an empty room, like, and you got to be continuously talking to that empty room until people start showing up. And once people start showing up, they’ll sit down, and they’ll start helping and listening or whatever. And that’s exactly how our community like our contribution campaigns goes. So it’s the most amazing thing to see a non open source versus someone that doesn’t come from that back. Because this is the epiphany I had, right or the experience I had. And they create these campaigns, they start creating all these help wanting tickets and asking for help. And it’s kind of like nothing, nothing, nothing. And then someday, one day, somebody shows up and does something right. And someone else shows up the next day, a little bit more, a little bit more than that. And before you note it before you know it, that person’s coming back. It’s like, oh, yeah, like, this is the most amazing thing in the world. Like, here’s all these things that I didn’t think people would want to do. But they just showed up and started helping, right? And for various reasons, right? Whether it’s their own interest, who knows why, but that’s one of those interesting journeys that they like, and then you see the convert, right? You see that person who didn’t come from the open source community, like, wow, that was a really amazing, helpful experience or whatever. And I just find that so fascinating and fulfilling, when that happens.

Matt Yonkovit:
Just to clarify, when you talk about a campaign, what you’re talking about is potentially like a feature or a spec or something that you’re gonna say, like, hey, we would normally develop that internally. But let’s try and see if we can get the help from the open source community. So basically, outline it like you would but outline it for external consumption.

Corey Hulen:
Exactly. Like we pride ourselves on being very open, right? In terms of how we try to spec it openly, we try to design it openly. We try to do all those things. So people can experience that in the open. And there’s, there’s campaigns that definitely take the feature route, like it’s a specific feature. And there’s campaigns that are just like, let’s call it, quality of life that are just like rinse and repeat campaigns, like, hey, we need to add test automation to the mobile app. Here’s like 50 tickets of where we could use help with automation right around the mobile app. And those are really nice because you don’t have to necessarily dive in and understand this huge spec in this feature. People love doing that. Don’t get me wrong, but there’s a subset of people who don’t as well. The other ones are really nice, too, because it’s like, oh, here’s a little bite-size thing I can do. I can just jump in and contribute in like half a day. Right? The other side takes days if not weeks or months. And I think it’s about having a balance of both those right because people love to eat. It’s like I describe it as like eating different kinds of meals, right, like you have your buffet. You have your fancy high-end dinner, and the reality is people I like both right I go to a nice buffet like having a nice fancy high-end dinner. So I think about for us, it’s about having that variety that you know that very sort of stuff or whatever.

Matt Yonkovit:
So are your engineers then building those help on it? Or do you have a community, people who build them for the engineers? Like, how does that work? How does that look?

Corey Hulen:
It depends on the types of campaigns that we’re running, lots of times our product management will help. The design will help. Engineers definitely want to log a large part of those leaders. Some of the things we’ve been trying to do lately is experimenting. It’s always easy in the and this is one of the things that perplexes me about open source, it’s always comparatively, it’s always 100 times easier to get an engineer to work on something than let’s say, a designer, or a QA person, or a documentation person, right? So we’ve been trying experimenting with a lot of campaigns around that like we had a doc campaign that we recently ran, I don’t know, probably a few months ago now. But even trying to do something external, we called like design crits and invited people from the community to come to help them participate. And I think we’re in our early stages of seeing how those things play out. And it gets me excited. But it’s also, there’s so much more to go there. Compared to engineering. I think everyone thinks engineering, and that’s great. But like, to me, there are so much bigger areas to talk about.

Matt Yonkovit:
Well, and I think that’s the thing that I’ve been preaching quite a bit about is that contributions are not just code. Yes. And so Corey without me putting words into your mouth that I can tell you my feelings on that, like maybe, maybe tell us like some of the things that you are looking for, from a project perspective that isn’t co-related, that would really help the team and help the project? Yeah,

Corey Hulen:
I think the big ones because there’s the biggest gap would be things like design, right? I would love to see more people in the open source community contributing to design, QA, whether that’s software automation, QA, whether that’s just manual QA, some sort of quality assurance, I think those are two big areas. Another third one is documentation. Those are all the areas that we’re kind of focusing on trying to focus on. There’s actually surprising there’s, there’s one other area that’s really popular, so engineering and co-contributions is by far the most popular. The second easiest, I would say for us as a company is translations like we have an amazing community of people who help us. Translators are always good. Yeah, yeah. And that’s another one that is just I don’t know what it is about that community. They’re just awesome. They just love translating stuff. I don’t know what it is. But it’s just, that’s another one where we do a lot of work there to help foster and grow that, but we don’t have to, we don’t really have to work as hard or nearly as hard. You know, that? I don’t, I forget what the coming language is right now. 16, or something like that. But a lot of that has I mean, 99.9% of that has literally been like open source community organic growth kind of stuff. Now we are on the company side. And on the projects, I put a lot of effort into it, right? It’s a lot of effort to maintain the ability to localize the software, and all that kind of stuff in the view. So we put a lot of engineering horsepower behind it. But the community comes in and does the translations for us. And it’s amazing. That’s one of those ones, it’s actually really good to be honest. But I’d say for me the other I’d be curious about your opinion, my aims would be design, QA and documentation would be awesome. We would love product management as well, product stuff as well. That was I think even harder, because it can be more nebulous in terms of what you want to do there. But I would love to experiment with something there too. But we’re doing a lot of experience around design and documentation right now. And it’s the documentation, one has been really paying off the QA one as well, I think they’re still really small, comparatively, the design one is where I think we’re doing a lot of iteration and trying. So we’ll see what happens.

Matt Yonkovit:
So for me, the ones that I have that aren’t necessarily yours, which they’re a bit nebulous as well, feedback. Right, like so and you can call that a little bit of product management, you can call it a little bit of QA, you’ve got a little bit of this as we get features in beta or that start making their way through the system. I have not met an engineer who doesn’t want to hear how it’s being used. And if people actually like it the way it is, right? And I hear that quite a bit where it’s oh, well, nobody told me whether this is used or not. So I don’t know. And so that feedback loop. I mean, you don’t think of that as a contribution per se, because there isn’t like, here’s a pull request for this. But participate in the discussions on whether it’s GitHub or Jira, or wherever you’re having those discussions. Oh, my God, that’s gold. Yeah.

Corey Hulen:
That’s actually another really good one, like feedback, like user testing, all those types of things, especially user testing on people who kind of know what you Like, it’s fine to use usertesting.com. But sometimes when they or something like that, but they don’t know what your sort of toolset is what you’re trying to accomplish, it can really vary, but I think that’s actually a really good one as well. Never really thought I haven’t even really thought about that one much. But that’s,

Matt Yonkovit:
yeah, yeah, the other one that I, that I am pushing, and this is, again, it depends on what the goal of your contributors are. You know, you talk about documentation and depending on where your documentation ends, will dictate, like, what is left, right. So a lot of projects have really good documentation on installation or setup, they have crap documentation, real-world usage, right. So the tutorials, or the how-tos, or this is how I did this thing. Those are things that are often missed, and like those, those are also like, their inspirational articles, if you will, or inspirational pieces of content. You know, I know, from like, a Mattermost perspective plugging in like other operations tools or other monitoring tools into that ecosystem you’ve got use cases and case studies where it talks about, like hey, this big bank did the security thing, they plug this into their Red Team Blue teams where they fought against each other to try and hack each other. And all this information started flowing real-time, all that kind of cool stuff. Like, that type of stuff. And how people did it is so valuable, because it not only tells engineers, how it’s been used in real life, it can then say, like, oh, well, it could have been 10 times easier if this thing got fixed, or if this other thing, and then another contributor might say, like, Can I do that for this other use case?

Corey Hulen:
I think that’s it. Yeah, those use cases, those practice templates have you want to scale and you are right, like, that’s, it’s really easy is relative, but like for us to get documentation on how to set up Mattermost With Nginx, or something like that, okay. Like, like, that’s a lot easier documentation to get than that, and what you described, which is like, here’s a scenario use case of how we really use it. And let us sort of giving back to that community but through documentation of like, here’s how you should, here’s how we set it up, maybe you can benefit from setting up and using it like this, like, yeah, that’s honestly, probably all borderline non-existent. Like we do that, from the company side, like we have our content, people who write that kind of information. I don’t know. And I probably get in trouble. I don’t know if I’ve ever seen something come in like something from a contract, excuse me, a contribution coming in like that.

Matt Yonkovit:
I think it’s hard. It is right, like, and I think depending on the software you’ll see some more than others. You know, me, we do infrastructure software on our side, right databases, and we’ll get some of those use cases, and they’re great, they’ll come and talk at conferences, we’ll have slide decks or videos. That’s great. And I consider those people contributors. And so what we did internally was kind of build a scorecard where it’s like a PR request that gets accepted as this and like if you do a blog post that talks about this, then it’s this. So we try to equalize that. So we can kind of figure out, like hey, you might not be doing code contributions, but you are still a super active contributor to just ecosystems.

You bring up a good point, I know, you’ve heard that before. And I always want to experiment there too, with exactly that, like sending community members to conferences and stuff and to speak on the project’s behalf and run if we don’t do that nearly enough. And that’s something I would love to sort of do, as well as kind of investigate that side. Because I think in the grand scheme of things, like we’re still a small company comparatively, the more we can have local community members show up and, and do something, like the better it is.

You have to admit the power of the mug is pretty powerful.

Corey Hulen:
It is pretty powerful. Is it users or open source people, the most amazing people in the world, right? They’ll work as home engineers especially, there are always two things that fascinate me about engineers, right? One is exactly that. They’ll free stuff like a mug like I’m in. Right? And the other one is, for some weird reason, engineers don’t want to pay for 99 cent abs, right? They’re like, Oh, I could build that in three weeks. I’m not gonna pay 99 cents. I’m like, Yeah, but nine cents. That’s like one that’s like five minutes of your time. Let’s just pay 99 cents, right? Even if I myself do that to a man like a Peter’s nine cents with that, yeah, you’re right. That’s one of those amazing things. And I think that’s actually another thing we do a lot of their communities we try to give them a lot of swag in terms of shirts and mugs and contributions and it’s it’s never gonna equate monetarily back to like the effort they put in but it’s a raw Hey, we really appreciate you that kind of stuff. And it goes, people forget, like that kind of stuff goes a long way because it’s more about the mission and those people are already there for a mission, right? They’re not there monetarily or anything like that, they’re there because for some reason There’s something they believe in, somebody wants to learn whatever that is. So they’re already there. They’re not doing it for money. They’re not doing it for these things. But I think the more we, as a project, and a company can say, hey, thank you, here’s a pat on the back, we really appreciate it and here’s a mug or a t-shirt or something in exchange, that those little gestures go a really long way. I think people forget,

Matt Yonkovit:
I remember having conversations with contributors that are at Mattermost. And they offer this mug and I was like, I’m getting that book, I’m getting that book, you know. And what’s funny is, a lot of times, they do kind of eclectic things, right? You just never know what people are gonna use your software for. There was one person that I remember talking with who used it to track the feeding schedules for his steaks.

Corey Hulen:
But we have, I think, three months now how many knobs are up here, but we have the original one you can get, you can get aboard monitor to be the focal board team. And we have actually a security bug and that even internally, people at the company are trying to get that one. So we have a mug, if you find a security issue, it’s like a black mug. I think when you fill it up with coffee, it is like lights or not lights up, but the words come out internally at the company. So the same thing, like if even if you work for the company, you have to make a contribution to one of those repositories before we’ll give you a mug. And like I said that the security researcher mug even internally at the company, people are like,

Matt Yonkovit:
Exclusive swag, exclusive swag, right. That’s always the best. But it’s great. You know, and I think that one of the things that I’ve always admired about the Mattermost ecosystem is that contribution base, it’s the fervor of fans. And you mentioned rewarding I know, one of the cool perks that I don’t know if you’re still doing but are you still doing Mater con and flying tribute?

Corey Hulen:
We did, but we didn’t do it virtually. And we did say we did it with Oculus Rift. And then we sent a bunch of community members Oculus Rift. So one of the things that you’re right, one of the so after COVID is over, we’re definitely getting back to that and having our matters where we invite community members and stuff like that, for sure.

Matt Yonkovit:
So just so everybody knows what matter con is, it’s the internal engineering kickoff, right, that you normally do. But some number of contributors become part of the engineering team basically for that week and are invited to whatever exotic location Corey dreams up.

Corey Hulen:
Exactly. Like we tried to get not just engineering, we tried to get the entire company there. But it’s obviously very r&d Focus, usually. And that’s exactly it, like, we try to split it into two halves, one kind of company half of the first couple days. And then it’s a really community-focused sort of all engineering man where we find a bunch of community members. And we just sit around and talk shop around Mattermost. And it’s a great time. And it’s yeah, it’s probably seasons, it’s usually a nice location and stuff like that. So it makes for a lot of these locations to make it better.

Matt Yonkovit:
So, always, always good, good to have those nice locations. Now. You know, it’s interesting, because I think that everyone hears, like what you’re saying, what I’ve mentioned, they’re like, oh, yeah, we totally do that. But I think the most difficult thing is, there’s, there’s kind of a leap of faith and some of this because from an executive perspective, and people who aren’t necessarily tied to the open source community, it’s hard for those people to see the value in some of these activities and some of the resources spent. So I’m not trying to call out people, executives, or investors, or anything like that. But what I’m curious about is, are there metrics that you found that are helpful, from a community perspective like hey, how do you measure the elusive ROI, if you will, of some of these activities? You know, I mean, certainly, I mean, I’ve used some of these as like, Man, I don’t know if I can, like, show the ROI on some of these. It’s a leap of faith that happy contributors are going to contribute more, and it’s better for all of us. But are there things that you’ve looked at or explored?

Corey Hulen:
Yeah, I mean, we tend to track different things, contributors, new contributors per month is one metric we really focus on. PRs from outside company, right from from from community is another metric we track and we, we pretty, we’ve tracked those pretty hard, like, we’re not necessarily in there every day looking at them, but we’re definitely reviewing them once a month, kind of sitting down as a whole r&d team engineering structure team was like, Okay, this is where we want it to be as it’s going up, is this going sideways? Is this flat and then trying to actively do those things that change that but I’d say those are the two things that we kind of focus on. Sure, there are a lot of others that are better. Those are just like the easiest ones for us to do. And they’re the easiest ones to see the most value because for me, there’s definitely a correlation in the sense that you’re right, like convincing executives at a company this can be hard, but it’s really easy when a bunch of like individual contributors at your company and staff members or employees at your company start saying the same thing, right? And it’s one of the and we see this like, well, we have high like, Oktoberfest with October, right is like an amazing month for us. I think it’s promoted for a lot of open source projects, they get a lot of contributions. You know, ours definitely spikes during this time. And it’s fun to hear and feel like at our company, especially in this month, like what’s going on, because we were like, This is amazing. Like, I can’t remember how one of the Reapers somebody was talking where it was, like 134 pull requests and more than 30 new contributors for just that one repository. Right in that one repository. Right. And, it’s fun to see the excitement as a company. And I think that’s probably not great advice. I think that’s one of the best ways is at our size as a company, is when you get everyone around the company really excited and seeing it, then that just bleeds up, right? If you’re an executive setting some of those functions, you’re kinda like, what’s going on? Everybody’s talking about this cool thing that’s happening, like, let me see more. But you’re writing it can be. I think it can be a hard sell sometimes, right? So if you don’t have open sources, your routes, or if you don’t have somebody at the exact level or whatever, at the investor level, who not only espouse it, but believes it, right. It’s one of those hard things, I think, and I think that’s kind of always the first key is just making sure there’s representation there. There’s somebody at that level, whether it be Ambassador level, whether it be exact level, board level, whatever, who has some belief in it, right. And it’s one of those. It’s hard, right? Because I mean it’s one of those mission-driven things, right, like it’s one of those things you have to believe in

Matt Yonkovit:
It’s a difficult thing. And I’ve talked to a lot of different people. And this is a struggle, right? So I get a lot of likes, so that’s when I talk with some execs, and I’m not talking about Percona is I talk with all kinds of people all over the place. And I talked with some of what you consider the stalwarts of the open source space. And a lot of them say things like, all metrics kill us because management changes and they come in and they go, Well, what’s the ROI for the thing that we’re doing? And it’s like, well, we’re building more contributors, well, what’s the ROI? So what? So there’s that elusiveness, right. And not everyone is actually looking for the same level of contributions. Right. So, yeah, that’s true a company that’s looking for contributions from a code perspective you might have a dev rel person or OSPO role, where it’s like, it’s all about the contributions. It’s all about code in and codes out. And you might have the same role, the same-named role, where it’s all about the adoption of the software. Yeah. You know, and so it’s weird because there’s a lot of inconsistency across the board when you talk about some of this. And there’s so there are struggles, right? So I often think like the people who are interested in adoption, they’re more on the ROI side. Whereas like, the people who are interested in the contributions are more on the ROI from, hey, look at how much work that the community did for us. Right, like, so they offset engineering, we hired two people, we did some stuff like this, it helped in all these different ways. But it might not help necessarily the bottom line,

Corey Hulen:
I, you brought up a kind of the interesting point, we call a free like a puppy, like you have to be very careful in the open-source world, right. And it’s hard, right? You know, it’s hard because like, we actively try to manage, like away from that. And the way we do that is we try to be very clear on like, here’s where we’re accepting contributions here, words where we’d like to go, if you’re doing something over here, that’s awesome. We want to support you and give you a tweet and give you a rah rah like forecasts go do whatever you want. But that’s probably not something for the core mainline. So those are hard conversations to have, right. But what I found is like, most people just don’t like, there’s kind of like three people, right? They’re there, they’re aligned to the project, and they’re working in that direction, people are awesome, don’t do anything, right. Then there’s the other extreme, which is they’re not aligned, and they have no interest in aligning. That’s the hard bucket where it’s and it’s small, it’s like, let’s say less than 10% by less than 5%. That’s the bucket, you kind of like, thank you very much like, you can go do your own thing over there, right. And the reality is, the majority of people sit in the middle, which is they want to help the project. And they’re just not aligned. But they don’t know that they’re not aligned, they don’t know that they’re working on things that like, it’s like, Hey, if you work on these things, the community would value these things so much more. And what we finally spend, the majority of our time doing is educating that middle bucket, right? That’s where we spend the majority of our time. And it’s a huge win-win. Because when I meet with most community contributors, oh, this is all I like, it’s like, it’s even in your real work web paper, I would much rather work on something that’s going to contribute to the company’s success, or whatever it is, or my boss is gonna see and be like, Oh, that’s awesome. But the same reality happens in the open-source community, right? People want to work on things that are going to be elevated, or that is important to the project, right? So everyone else can see it. It’s no human ego in some respects, but And so we spend a large bulk of our time just educating those people in the middle of like, hey, that’s awesome. We want to support you. But like, if you’re working on this other thing over here, like that’s what the community really wants, like, like the project really? Oh, that’s okay. That’s yeah, sure, I can do that. That’s awesome. I don’t really. I’m not married to this idea, kind of attitude. And so a lot of that happens, too.

Matt Yonkovit:
Yeah, no, I mean, it’s, it’s an interesting space because the community can be so open and so involved, they get so passionate about certain things and sometimes it’s hard to have those conversations, because like you said, there might be that 5% of people you’re like, Wow, that’s great, but like to maintain what you just did would like cost an entirely new engineering team, we would have to double our staff. So it’s always difficult but you have to align things externally and internally so it’s cool advice to see how you’re doing those campaigns and the help on it and I would encourage other folks in that space to take a look at how they might be able to replicate that. So product-wise for those of you who don’t know what Mattermost is you probably see it soon, and so we can we can call it probably do a disservice by saying that most people think of it as chat.

Corey Hulen:
Yeah. persistent chat like Well, we like to consider ourselves communities to our platform, but topic base or channel based but our sort of if you know what slack is we’re like an open source equivalent to that alternative to that. I think the more interesting thing is like, where’s the space we play? And for us, it’s definitely around privacy and security, at least when we separate the thing. There’s the open source project, but the company side of it, obviously we spend a lot of time in that sort of secure space, that privaсн space, that’s where our market is. And that’s where we tend to do really well.

Matt Yonkovit:
Yeah, and, but, I mean, it’s grown beyond just chat. I mean, we’ve seen this with other like, like communications platforms. You know, a lot of people are starting to integrate their pipelines, they’re DevOps automation. And so, it’s chat ops. So maybe tell us a little bit about what chat Ops is, and like, how that’s evolved?

Corey Hulen:
Yeah. So I mean, basically, I mean, chat Ops is probably an older term now. But it’s exactly like taking your CI/CD pipeline.

Matt Yonkovit:
Okay, like, educate me.

Corey Hulen:
And that’s a good question that I think it’s more about just we call it developer collaboration, but I’ll describe it and implement products that we have, because there’s that whole notion of taking a lot of exactly that, like your CIC pipeline, your automation tools, your tools to deploy stuff to the cloud and integrating those into your chat application. Right? We have, like I said, one of the things that we’ve been investing a lot of time in is what we call developer collaboration, right? So anything where a developer or and we use the word developer very loosely, it’s not to say somebody who would go, who writes code, but anyone in that kind of knowledge worker kind of space, right? Who needs to integrate those types of things. So that’s where we’ve been really focused as a platform. So we have Mattermost channels that this communication does at the day, we have a playbooks feature, which is basically incident management or incident response. So if you think of things like checklists, or what’s typically called SRE incident management, so you have an outage, what’s the checklist of a playbook or the flight you run to bring that outage backup stuff like that. So we have a whole mechanism or toolset, application around helping you manage that kind of workflow. And the most recent, I was really excited about is actually boards. So Mattermost boards are completely integrated matters. And that’s basically a kind of style board. So project management. And we’re really trying to focus on developer collaboration. So what are those things that you do as a team where you need communication, collaboration, and to work together as a team, that’s where we consider our sweet spot is around all of that sort of people side of stuff? Yeah, you still have your C/CD tools, your GitHub actions, or GitLab, CI/CD, whatever it is to run those types of things. But that’s where we see ourselves being very strong is sort of helping people with what I call unstructured data, right? You have your structured data coming out of the source system, it alerts, it pops into Mattermost, then you have some other process that responds to alerts. And that process, at least in our world, is always very unstructured humans talking to humans like you call them war rooms, whatever it is, right? You have an outage, everyone else in a war room, and stuff like that. And the reality is, it is a very unstructured process in terms of discovering what the issue is, but it’s a very structured process in terms of how you go about that. And so we’re trying really hard to help with that use case, as one of the use cases around that sort of incident management incident response as well. So that’s kind of where we’ve been playing a lot. Like, I think developer collaboration is probably the best like the funniest one is when you go talk to somebody, people who do like really large Incident Management, like really big like, like national security level types of incidents, right? And it’s kind of funny, right? Because you talk to them, and they have these alerting tools, these security alerting tools that are popping and going off. But the reality is, lots of times they have those conversations in Mattermost, like the people conversing about are actually in Mattermost, or in Slack or something like that. They’re in some sort of chat topic-based tool. And that’s where the real work is happening, right? Yeah, you have your alerting systems that fired it off. But that’s just kind of like the start of the workflow right at the start of the workflow if an alert happens. Is this important enough? Like or is this a false positive, important enough, then it gets piped into Mattermost. And then sort of that process takes over from there. And we see a lot of success around that because that’s really the reality of it is helping those teams sort of communicating, collaborate, stuff like that. That’s where our sweet spot really is.

Matt Yonkovit:
And I mean, that could be a lot of compliance help, too, right. So if you’ve got everything in one location,

Corey Hulen:
yeah, exactly. Yeah, compliant, I mean, security, at least in our world security and compliance go hand in hand. As I said, we’re in a lot of large banks. I think three of the five largest banks in the world run matter most. And obviously like now he says security and issue but like company clients is really an issue in those worlds, right? Because you have things like insider trading laws, where they’re now liable for even 10s, or hundreds of millions of dollars and all those types of things. So you and that, that’s one side of the coin, the compliance is one side of the coin, then we also do well on the other side of the coin, which is the pure privacy side of the coin as well. So that’s kind of like two different sides of the privacy coin, what is the business or whatever is required by law to regulate, like what they’re doing? Right? And then there’s the other side of it as well. So we played well, in both those areas.

Matt Yonkovit:
Yeah. And you keep on adding more tools to that toolbox to make it easier to kind of go back and forth. Because I mean, I think that, ideally, when you’re in the soup and you’re trying to deal with one of those crazy outages, having access to access data or tools directly and having outputs and things is important. And I know, you just launched Mattermost 6, which is a news release. And so you have a new notion feature. Yeah, maybe ever.

Corey Hulen:
Yeah, we combine some boards as integrated restraining anonymous, links back and forth between and says exactly that. And that’s, that’s what we find lots of times, right people like context switching, especially in that kind of environment where seconds or minutes cost 1000s or 10s, of 1000s of dollars is really important to sort of be it’s honestly, it’s not even just for those types of outages, it’s just like a developer quality of life issue. Right. Like, customer like I call it customer satisfaction, but like developer customer satisfaction issue, right? It’s like, oh, great, I get this is like the 14th like a disparate tool that I have to go log into to find something buried in the basement somewhere. And nobody’s going to fix that problem. But I think everyone’s working in that direction. And there are the quality of life, things that you can do, right, you can make it a lot easier to turn a light switch on, Oh, this one has a staircase down in the basement, all those types of things. And that’s where we really find ourselves focusing on that kind of context, switching quality of life, we think there’s a big a lot of opportunities.

Matt Yonkovit:
So I mean, other than the quality of life issues, like are there things you’re hearing from the developers or the engineers like that these are the things that they’re worried about, or concerned about right now?

Corey Hulen:
Yeah, that’s a good one. I mean, I, I think security is always one of those worries, I think that’s one of the things that we always hear about, and the more we can integrate with security tools, tooling, that’s always a good spot on in compliance, it really depends on if you’re in the ops side of the house, or the dev side of the house. So if you’re on the ops side of the house, it’s like security compliance. If you’re on the development side of the house, it’s actually things like keyboard shortcuts are quality of life in the UI in terms of performance, like, it’s the little things that actually score the highest, right, and that’s what it is, if you don’t it’s one of those things, it’s a move you do by 1000 times a day if you or whatever, 5000 times a day, and you save half a second or point by a second, like point two, five seconds or whatever, that’s, it’s a big deal. So we get a lot depending on like, if you’re talking to the dev side of the ops side, it’s very different feature sets, very different asked, but I think still very valuable, because the reality is, are those two worlds DevOps will now even using the term one dev SEC ops, right, they’re kind of merging all three together.

Matt Yonkovit:
Hey, we’ll break them apart, merge them continually over the next 20 years. But it’s interesting that you mentioned, like some of those, those small things that matter, so much Mattermost. But that No, no, sorry, that was a total pot. But it goes back to that kind of UI UX focus, it goes back to who your audience is. And it’s funny because you mentioned and we run into this too, right? Where, hey, from a database perspective, most of our UI is a shell. Right, like kinda is, right now, as more developers start to manage databases, then that changes. But it’s interesting when you talk to people who have that, that straddle that line, right. So you’ve got deep operations people who may be comfortable in scripting go into the shell doing some command-line operations, and then you’ve got all of these users who might be consuming. You know, what, you’re what you’re giving. And that UI UX for them. Make sure to breaker adoption. Yeah. Right. because I mean, let’s be honest you deploy this at someplace, and let’s say there are 10,000 users, probably only 100 of those users are those deep operations people who would actually like to be able to do the command line. So you’ve got to do that. 9900 people, you got to make them happy. Yeah.

Corey Hulen:
No, and that’s totally true. Like, especially for us around sort of communication collaboration, your spouse On, right, like, we have to make those people happy. We have to do it from a very I think slick and clean and quick UI UX perspective. But it also raises an interesting question. And I know this is changing a little bit. But if you look at a lot of the previous successful open source companies, a lot of them are infrastructure. And you’re right, right? Like, they don’t have to invest in UI, like, like we do as a project. And as a company, it’s one of those really hard things sometimes because you don’t get a lot of those natural, open source resources. Like I can’t say we try to try and increase the design. But even finding somebody who’s, who’s even fine. I mean, people do it because they want to experiment, but finding someone who’s truly passionate about like, react, native open source, and is contributing to Mattermost, or whatever, like, it’s definitely not as easy as finding someone to contribute to the backend server or stuff like that. And I think it’s one of those interesting things. I don’t know why it’s like that. But it’s definitely, anecdotally at least you look at a lot of the big open source companies that are successful and blah, blah, blah, it’s a lot of it is infrastructure pieces. And I’m sure we can go into detail about why and how but it’s just, I know, we always find that to be challenging because we’re one of the few sort of or frameworks. There’s a lot of frameworks out there. Right. But we’re one of the few, I think, true applications where we have end users, right, using our application, like you said, right? They’re not people on the command line, there are people who don’t even know I would say some large percentage of our audience doesn’t know what that is, right? But the reality is, if we have to cater, I always like to describe it. And maybe this is my own sort of ego coming through, I always like to describe it as like, you’re at a nightclub, right. And like, all the developers are inside partying, and there’s like, there’s the velvet rope trying to keep everybody else out because the developers just want their own little room to go partying. But the reality is, it creates a FOMO. So you can join everyone else at the company who wants to go join the party. And that takes the form of things like chat, or whatever. And we actually see that lots of times with our customer base, where that exact same thing happens where they spend a man on the server, it’s meant just for the developers, but it’s like Oh, do we need this person from design to come in? Okay, finally, we need this person. And it gets in like, we really need this person for marketing, because we know there’s this message that has to go out. And so what ends up happening ends up becoming the space that everyone just organically ends up morphing to because of the developers there. So that’s sort of our mantra, like, we focus on developers, that’s our tip of the spear. That’s the audience we want to own. But we have to pay homage to those general end-users, because you’re right, like, they will just, it’ll be like organ rejection, they’ll just like, No, I’m not gonna show up there because I gave you this application. And it’s, it’s a hard balance.

Matt Yonkovit:
Well, but I think that a lot of infrastructure projects in the open source space need to learn from that, because we’re flipping this power, right? So when you say like, oh, you’re focused on the developers, most of the infrastructure folks have been focused on the SIS admins DBAs, that’s our use maybe. And now, those people are building internal platforms to do their kind of infrastructure as a service, or infrastructure code external, or to their developers. So even if your core audience has classically been the back end, now all of a sudden, you’re moving one step closer. So it’s changing how a lot of perceptions are because what used to work doesn’t necessarily work. And you have to think, step up.

Corey Hulen:
And I’m so excited about that feature where exactly that like more design, more front end work is coming to open source because you rewinder and your typical open source front end are like, you open the Properties menu, and there’s like this huge page with like, 50,000 checkboxes, things you can turn on or off. And that’s great if you’re a nerdy developer, right. But if you’re not a nerdy developer, that’s just overwhelming, right? And so there’s that great balance that has to occur

Matt Yonkovit:
If you want something that’s gonna say, like how security you want this, my grandmother can get in like the people next door with the tinfoil hats camp. You think whatever I mean, like, you make it so it’s so easy. The three buttons and you’re done, or whatever, I get it. Right. And I think that everyone wants it easy. Yeah. And they want it as fast as they can without having taken that. And you’ve designed for that. And I think you’re pretty good. Yeah. No, that’s true. Yeah. But Corey, I wanted to thank you for coming out today. We’re almost at the end of our time, I want to be respectful of the time here. I appreciate you sitting down talking to us a little bit about the Mattermost community, some of the cool things you’re doing. You know, we always want to inspire other communities and folks out in the open source space to learn from one another and share these topics. So it was great to have you on.

Corey Hulen:
Awesome. Enjoyed it. As I said, love it. It’s always fun. It’s always fun talking with you, Matt.

Matt Yonkovit:
Thanks, Corey!

Did you like this post? Why not read more?

✎ Edit this page on GitHub