David Cramer - CTO at Sentry

(Joel Beasley at 00:00:00) Hello, my friends. Today we are talking to David Cramer, the CTO and founder at Sentry, and we discuss his transition from CEO to CTO, how their software monitoring is helping enable productivity, and the biggest lesson he's learned as an entrepreneur and founder. All of this right here, right now on the Modern CTO podcast. Here we go. This is the Modern CTO podcast.

(David Cramer at 00:00:33) David. Hey, Joel. Good to meet you.

(Joel Beasley at 00:00:36) Good to meet you, man. I'm pretty excited to talk with you specifically because you managed to take an open source project, build an open source project, and then build a business behind it. I think that's pretty cool.

(David Cramer at 00:00:47) Yeah. Lots of, not even hustle, just lots of hard work.

(Joel Beasley at 00:00:51) How long did it take from the time that you first had the idea to do the open source project until the open source project actually existed?

(David Cramer at 00:01:01) So I don't know if this is kind of traditional for open source, but, you know, Sentry wasn't planned. It's actually probably very traditional in open source. You know, I think you kind of have founders and they're just like, I'm going to build something and they just spin up ideas and they're like, okay, this is the one I'm going to build. Open source, a lot of the stuff is just born out of like, oh, I need to solve this problem right now. That's no different. For Sentry, it actually started as this example piece of code that somebody else wanted to understand how to do. Think of it like Stack Overflow, but, I don't know, twelve years ago. You know, it might have existed then. But, yeah. So somebody asked how to capture these errors into a dashboard. I'm like, that's kind of interesting. Why would you need that? And I don't think you so much needed it twelve years ago. But yeah. So that started, sort of the project, and I had a lot of open source projects. Basically, everything I did was open source because I didn't see any reason for it to not be. And Sentry was just one of them that kind of kept growing with me and I kept investing into, and I don't know, just a very long, continually iterative journey of that investment, I guess, has kind of grown it through different stages ultimately to where it is today.

(Joel Beasley at 00:02:14) And so did you get some early contributors that really helped keep the momentum going?

(David Cramer at 00:02:20) So I think for me, you know, I think different people get excited by different things. Like, you know, building things was like, building software is always my hobby. You know, I'm self-taught. And having people use the software is kind of the thing that drove me. And developer tools are like having your peers use your software, so it's even better. And so we had some contributors, but frankly, because of the way Sentry works, most of the contributors ended up contributing sort of integrations with third parties. So something like Slack, but imagine Slack twelve years ago, or integrating with new programming languages. So like, JavaScript, the original implementation was contributed by, I believe, somebody that was working at Mozilla at the time. So that's where a lot of it came in. But most of the core server I actually built myself and continued to, you know, hack on, I guess, in my free time.

(Joel Beasley at 00:03:12) The name is awesome, Sentry. Was that the original name?

(David Cramer at 00:03:16) So the original name was Django DBLog. It was created—so Django is a web framework, and it was a database log. Like, at the end of the day, it was very, very pragmatic. But no. So I joined a company called Disqus. They're not around so much anymore, but, and they were using it, and it was not very good. And we had to stay—I think StarCraft 2 was about to come out or had just come out or something like that, and so all of our projects, we were just naming after StarCraft units. And you tried to say like, what unit sort of resembles what you're doing right here, and Sentry is a little bit of a stretch. But, yeah. I mean, we ended up just renaming the project, like rebuilt it sort of at that point. And I think that's really where the foundation of Sentry came into place. Like, it did exist before that, but that's really where I spent a lot of time. And that sort of marked a three-year journey of all nights and weekends kind of hacking on it, and it had grown significantly by that point too.

(Joel Beasley at 00:04:19) And then what point were you—when did you make your first dollar?

(David Cramer at 00:04:23) I have a calendar reminder for this because I can never remember. I think it was in—so it's 2020. Let me see if I can even—I think it was 2012-ish. Probably early 2012. I'm going to—I might be off by one there. But, you know, we spent, I think, two and a half to three years with it running as sort of a side project where, you know, we're just generating revenue and stuff and using that to fund, you know, the server bills as well as fund various sponsorships of third-party events and conferences and stuff. But it was very few dollars at first, I would say. We charged nothing, which, you know, is good and bad.

(Joel Beasley at 00:05:05) I like the brand a lot. You know, I was checking out the website. You have a really—I don't even know how to describe it. It's like midnight cartoon or something. And I don't know. What would you describe that as?

(David Cramer at 00:05:19) I would describe it as don't take yourself too seriously. Yeah. We put a lot of—so my cofounder, basically, the person who helped me start the business was the head of design at Disqus. And so I convinced him to come along mostly because I super value good product design, and I can't do it that well myself. Like, if I have to design something, it's just every shade of gray you can imagine. And so I roped him in, and I think I kept that foundation of really, really caring about design. And it is certainly hard to keep that up, and I won't say we're perfect, but we invest a lot into visuals and to the point where one of our core values, we say pixels matter. And it's in the most narcissistic way where it's like, call out every minor detail and challenge the minor details because it adds up. So if you try to change something in the product and the margins are off, you'll have five people instantly call you out. And I think we actually think that's super important because if you think about brand, and we care a lot about brand, the first version of brand is visual at the end of the day. And so we've got a creative team that runs the external brand which you see on the marketing website, which is actually run by a former GitHubber who, I think his claim to fame is the GitHub icon, which they still use today or at least a variation of it. But then we've got another team that runs product experience as well. But it yeah. It's a super important investment for us.

(Joel Beasley at 00:06:48) Yeah. I'm a big fan of branding and design, and it's because when I interact with other products, the moment I land on their website, I form an opinion on their brand. And it gives me instant insight to what their product will be like. Because it's just like, these are the people that work there. You know what level of designers they have because a good designer will come in, and if one part of the business is off, they'll deem this unacceptable and raise it up to the standard. And so you can kind of—you could tell so much about the collection of people by the products that they produce.

(David Cramer at 00:07:23) Yeah. Yeah. I do think it's really a representation of what you care about or if you care at all, I guess, at the end of the day. Like, ours is a representation that we really care about the results of our product, and we've always been product-first. And so I agree with you. I kind of look for it, and if, frankly, a product or a company has terrible design, usually, the product also ends up reflecting that.

(Joel Beasley at 00:07:47) So I saw that you made a move from your CEO, founder, but then you moved to CTO. Now recently, I brought in a VP of sales, cofounder person to run the sales organization and to handle that part, and that changed everything for my business. Like, I'm a lot smaller than you. We're only eight people right now, but those initial people matter so much, right, to develop a solid foundation. And it just felt like a huge weight off my chest. How are you feeling after this transition?

(David Cramer at 00:08:18) I feel great. Yeah. So the way I always tell people this is, when I first raised money for the company, which is, you know, about five years ago now, one of the first questions that was asked of me is, like, do you want to be CEO? Which I actually think is a really, really important question to ask a founder because being CEO is very different than being a founder. Right? And I can't remember this clever quote, but my answer was we'll see. And it reminds me of this movie where there's, they're telling a story about some monk or something, and his answer is always we'll see. Like, we'll see what's next kind of thing. And that was always my answer every year when either somebody asked me of it, if I wanted to continue that role or if I asked myself of it. And then, you know, early last year, like at the start of last year, I asked myself the question again. I'm like, yeah, I don't think I want to be CEO anymore. In a sense of, the job was very much CEO. It was no longer founder. It was no longer just get it doing anything, hack on stuff. It was very much hire executives, manage executive teams, set strategy—and some of it was fun, but frankly, hiring executives is my least favorite task. Like, it's—I don't know. It's kind of one of those things—I'm a believer in do the thing you're great at because you have the most impact, and I am not great at running a go-to-market organization. I'm not great at, you know, doing executive sourcing or any of this other stuff. I think I'm really good at product intuition, and I'm very technical. So I'm like, you know what? Let me see if I can find somebody that balances that out really well, takes on more of that executive side. It's been really great so far. Like, our core executive team is basically, you know, we have myself, my cofounder, and then I hired a CFO maybe almost two years ago now, and then now the CEO. And that team's actually been really great for a sounding board, which, it's kind of one of those things that for me until I experienced it, having a couple seasoned executives who were very trustworthy and you, you know, you could just guarantee they would execute on whatever they said. I don't know. For me, it—so I would agree. It was very much a game changer and still is. You know, we're six months, almost six, into having Milin on board and still going well. So, you know, I don't envy him right now having to deal with the sort of current climate, but he's doing a great job.

(Joel Beasley at 00:10:35) Right? Yeah. There's a couple CTOs I've been talking to that have had new CEOs towards the beginning of the year or back in October, November. And I'm just like, oh, man.

(David Cramer at 00:10:47) Yeah. Yeah.

(Joel Beasley at 00:10:48) What a tough moment to walk into. But I'll tell you what though, it will really set the tone for the next decade. Right? Like, you'll really get to understand who this person is and how they think and their capabilities. And you don't have to do all that communication. I mean, you have to do the communication to your team, but you don't have the responsibility of communicating to the entire org and all the investors. I like it. I definitely agree with you—find what you're really good at and what lights a fire inside of you, and then spend 80% of your time there.

(David Cramer at 00:11:19) Yeah.

(Joel Beasley at 00:11:21) Okay. Dude, this is still early. I haven't done a podcast this early in a long—I don't think ever. I've done some late ones when we have to do Australia and stuff. But it's interesting how my mind's in an entirely different gear. How are you feeling?

(David Cramer at 00:11:35) Better now with coffee. I'm not a great morning person. I have gone up a lot earlier, consistently since, you know, shelter in place, which has been, I don't know, good. But I feel like I just don't know what to do with the rest of my day. Like, 2 p.m. comes around and I'm like, I feel like I've done everything I'm going to get done. So now what kind of thing? But, yeah, little early for me. I'm assuming you're on the West Coast as well.

(Joel Beasley at 00:11:59) No. I'm on the East Coast, but I usually do—I mean, it's only 10 a.m., around 10 a.m. here, but I usually don't do podcasts until the afternoon. So my mornings, I get up pretty early. Usually, I get up around five or six, and then I run, and then I make breakfast, and then I go lift weights, and then I, you know, come into the office and do all of my most important tasks first with the team and everything. And then I spend the afternoon in podcast social mode, you know, because that's 2 p.m., and it's like, alright. I've done everything or, you know, let's have some awesome conversations and things like that. So, yeah. It's just different, but that's the variety of life. Right?

(David Cramer at 00:12:45) Yeah. Yeah.

(Joel Beasley at 00:12:47) Do you get into any of the performance stuff or working out or personal development?

(David Cramer at 00:12:52) I got into backpacking for quite a lot of time. And I just—I don't know. Past couple years, it was hard to find the time to do it. And this year, it's not looking good just with park closures and everything else, unfortunately. But, yeah. We'll see. I was hoping to get back into it, but, you know, the summer looks like it's shot at this point.

(Joel Beasley at 00:13:13) It's weird how they can close the outside. Right? It's a weird concept.

(David Cramer at 00:13:17) Hoping that—yeah. I'm hoping to open some stuff back up.

(Joel Beasley at 00:13:20) Wait. Is stuff still closed there?

(David Cramer at 00:13:23) Some of the stuff is. Yeah. So I think it's case by case. I think a lot of the national parks are still closed in the sort of California stuff. It kind of depends, but we'll see.

(Joel Beasley at 00:13:35) So walk me through what backpacking is to you.

(David Cramer at 00:13:40) So I think what I really enjoy about it is it's like a forced disconnect. I hate disconnecting for what it's worth. But I really enjoy the outdoors, and if I don't have a choice but to disconnect in the sense of there is no cell phone service, it's super easy. And so for me, I actually just very much like the camping part of it, but I appreciate the fact that I have to hike in somewhere semi-serious and take some time to get there. It feels a little bit more remote or a lot more remote in some situations. I don't know. Just a good escape. I always do it with friends, though. You know, I'm a social person. So but, yeah. I don't know. That escape. It felt like it was a good balance. It was a good way to—I've never been a gym-goer, so it's also a good way to keep in shape. You know, I coupled that with deciding if I was going to do more climbing outdoors, which I think I managed four or five trips. So I did not get very far along that endeavor. But, yeah. I don't know. It's a really refreshing thing to do. But I also think it's become super mainstream lately in the sense of, out here in the Bay Area.

(David Cramer at 00:14:49) There's a bunch of cool parks nearby, and Tahoe is a really well-known area. And when I started, you know, this was like, it's gotta be like six years ago or something. When I started backpacking around here, you could sort of just with a couple weeks notice go get a permit, and it was great, like an overnight permit. And now, like, the whole summer is basically booked up instantly, which, you know, kinda sucks. But yeah.

(Joel Beasley at 00:15:16) So it's like a permit to go camping?

(David Cramer at 00:15:18) Yeah. It's a permit to hike in and backcountry camp overnight just because they wanna keep it manageable, clean, so they only permit so many backpackers in each evening. And it's pretty common here.

(Joel Beasley at 00:15:33) It's like a byproduct of like a bigger city or a larger population because I live in kind of a rural area. So there's like, there would be no permit or anything. But I could see how if you have like millions of people in San Francisco and then you have like a couple nearby parks that like everybody can dump out into them and just completely overrun it.

(David Cramer at 00:15:53) Yeah.

(Joel Beasley at 00:15:53) You need a permitting system. That makes sense.

(David Cramer at 00:15:57) Yeah. It's good.

(Joel Beasley at 00:15:58) Do you do any gaming at all? Are you into gaming?

(David Cramer at 00:16:01) Yeah. Very much so. Like, ever since I was a kid. I actually get a lot of my start with building software because of games. Like, early on, I was, I got really into MMOs. I feel like partially because it's when the internet was just like growing and becoming big. Like, I played a couple of the first online games, and then I would just build basically like content databases out of those games to the point where I started learning to like reverse engineer those games and sort of automatically populate the database and stuff like that. But these days, I'm still into it, but it's more like, you know, some people watch TV, some people read books, some people do whatever for entertainment, and like, games are that for me. But they do not hold my attention like they did when I was a kid. That's for sure.

(Joel Beasley at 00:16:48) Me either, man. I actually when this whole virus thing hit, I went out and got like an Xbox because I've been, I was gaming like my whole life hardcore from like, I remember sitting in front of the TV waiting for Xbox Live to come on. You know? Like, I think I was in high school. Right? Because there was like a big countdown for when Xbox Live would turn on. And I played all types. I played consoles. I played computers, everything. And so I remember, you know, going through all of that, and I would just, I would play like just all day, like all day. And then now it's like, you know, go in there and just rank in the top for two games, and then I'm good. Couple games. You know? It depends on because for me, honestly, like, I've built up my life enough to where being here and interacting with people is more rewarding for me at this stage.

(David Cramer at 00:17:40) Yeah. Yeah. I think mine is, it's been different since shelter in place, of course. But prior to that, it was like, once in a while, I would take like a Saturday and I would just like game or kinda do nothing. It was like super downtime, which honestly for me is great, kind of a way to recharge. But it just, especially now with shelter in place, I'm like, I don't know. I can't handle it. It's just too monotonous now.

(Joel Beasley at 00:18:05) Is most of your employees, are they like in San Francisco with you or they just spread out remote?

(David Cramer at 00:18:10) Yeah. We've, we do distributed. We don't do remote. So we are primarily here in San Francisco, but we do have two other offices in Austria, in Vienna, and in Toronto.

(Joel Beasley at 00:18:21) Okay. Cool. What made you choose those locations?

(David Cramer at 00:18:24) So Vienna was, I'll say it was an accident. One of the first folks that joined the company lived, I don't know if it's in Vienna at the time, but in the area. And I thought for sure I would convince him to come over to San Francisco, and I did not. And so we're like, we'll just spin up a team there. And then that actually worked so well that we're like, you know what? We actually have some leads in Toronto. We know some of the folks. You know, let's do the same thing there. So Toronto is still smaller. It's earlier. But Vienna, I think, is around 20 people now.

(Joel Beasley at 00:18:53) Were your first couple hires like salespeople or engineers?

(David Cramer at 00:18:57) I think the first, you know, 15 hires were engineers. And that roughly stayed, you know, other than one here and one there, that was like true for a very long period of time.

(Joel Beasley at 00:19:07) And so were you just getting a lot of sales to people like using the open source product and then wanting, and then that's that was like your pipeline for growing your business?

(David Cramer at 00:19:17) So I would have a hard time sort of saying what the pipeline was, but it was very bottoms up and very much like, you know, sort of field of dreams-y. If you build it, they will come. Like, we just had, I think we had a good reputation or have a good reputation and a good brand, and people like the product and it was affordable. So just my top of funnel just existed, and a lot of it was word of mouth. And so I think open source helped with that. It helped grow brand over time, but, frankly, the people that were buying Sentry SaaS were not the same kinds of people that wanted to run Sentry open source. It's just like a different persona half the time, and we learned this because we realized many people buying our SaaS service had no idea we were open source. They just knew Sentry was a great product. Right? So that actually has done really well. And even today, like, most of our revenue and it's slowly shifting, but most of our revenue comes through the self-serve channel where it's just people sign up, pay with their credit card, which to me is the best kind of business you can build. It's super validating. It's also super sustainable. But we are shifting a little bit more where there is a traditional sales org, and that's actually become valuable for certain kinds of customers that are used to that and expect that. But yeah. I don't know. I think we were very fortunate that we were able to just hire a lot of like engineering personas for a long period of time, and we didn't have to have a traditional sales team in place.

(Joel Beasley at 00:20:38) Yeah. Especially when it's not your strength or when you don't have experience in like building those things. For me, I don't know if you know a lot about my background, but I've been programming for about 17 years. And I built a number of different softwares. And usually, what I was doing was I was building prototypes and then selling them off to companies, or I was working with venture capital companies to help rewrite systems that were getting funding so that they could scale. But then I, you know, started my own thing, and I didn't have a lot of sales experience because those businesses, the business just kinda came. It was just kinda there. And to build a B2B sales team and to push that out there is like, it is a massive challenge. I mean, it's basically like doing something completely new that I had no experience in doing. And it took me like seven failures to find, to figure it out. To figure out like who the right, like, what does that human look like? Like, I could hire engineering teams and build those structures all day because I know what those people look like. But to know what a good salesperson looks like was really tough. And then you know, you can have good engineers, and then you can have good engineers that could scale teams. Right?

(David Cramer at 00:21:45) Yeah.

(Joel Beasley at 00:21:46) And so to find those great people was, it was a lot of work. How do you, like, do you have any tips, or have you learned anything about finding great people?

(David Cramer at 00:21:56) You know, I think it's tricky. Like, I've also been through situations. Like, I think there's a lot of great people out there, and I think the mistakes I made when hiring folks was not finding people who were right for the company or sort of right for the time of the company. And my view of the world has always been now, like, get to know the person much more because you can only trust so much about their execution from what you can, like, engineers are super easy to evaluate. Like, you can come up with a rigorous curriculum for exactly, like, are they gonna be able to solve this task? But you can still hire them and they'll fail. And my version of that, like, at a startup especially, would be like, maybe they're not like, they don't have a lot of self-drive or, you know, they're not a very hard worker, like, work ethic or something like that, or they can't figure out work-life balance. So like, there's lots of ways they can fail even if they pass that test. But at the very least, you can validate the test. Right? Like, can they solve the problem? When you're hiring like sales and you're not a salesperson, how do you evaluate that? And then for us, it's like, when you're hiring sales and you have a bottoms up business and the buyer is often a developer, how do you evaluate who can sell to that person? And I'll let you know when we figure that out because that is like really, really tricky. But, you know, we basically just, I think the biggest lesson I learned as a founder and especially my time as CEO is just like, trust your gut. And like, even when I talk with our current CEO and he's like asking my feedback, I'm like, I have no idea, but like, I'll work through it with you and like, just like, if you don't think it's good, don't do it kinda thing. Like, it never works out well when like something's telling you it's a bad idea. And, you know, often that comes to hiring. But yeah. Because I think every situation where it's like, it felt like a roll of the dice very much was a roll of the dice, and you can see it in hindsight. And I don't think any of those worked out. And so we went through a lot of challenges in figuring out how to build a sales program. And even these days, like, our marketing program is basically brand new. You know? We're 120 people in the company, and we have, I don't know, maybe four or five in marketing now. And like, we don't do a ton of campaigns, and the sales team is probably 15 all in. And so I don't know. I think on the counter side, like, use your strengths. You know, we knew we weren't good at sales or didn't know sales, and so we always focus on how do we solve problems through product. How do we build a better product that can solve that issue for them? But, yeah, I don't know. When it comes down to like finding folks who you do not have the skill set you're hiring for, I just think it's really difficult.

(Joel Beasley at 00:24:34) And then can you give me like the 30 second, like, pitch of like what Sentry is and why people buy it?

(David Cramer at 00:24:42) Yeah. So this has evolved over time, but I always describe Sentry. I give up on, you know, simple big slide sentences, but like, I think about it. It's like you load up the Uber app, it crashes. We tell Uber's developers about that. And we think that's important because as a customer, all you do is load up Lyft as soon as Uber crashes. You don't actually care. There's no brand loyalty at the end of the day. And so, you know, it's bad that you as a customer experience that issue, but it's really important that the software team can understand and address that issue. Right? And so that's how we started. We call it crash reporting, error monitoring, something like that. But these days, like, a lot of what we do is sort of expanded where that same idea of like, we just wanna tell the developer about like this defect in their application. We're just continuing that, but we're identifying more types of problems. And so I explain that analogy where it's like, okay, you load the Uber app, it crashes, obviously bad. Like, we can all recognize that. But if you load the Uber app and it sits at a loading indicator for 30 seconds, that's just as bad as it's crashing. Or if you load the Uber app and it's like putting you at a login screen, but there's no login button, so you can't actually do anything, just as bad as the loading indicator, just as bad as it crashed. So we think of like, there's all these different problems, and we just wanna be able to identify them, get it to the software team, the product team. And that's it. That's really about like the efficiency of being able to achieve that. And I think the difference or one of the big differences in Sentry versus, you know, logging or any of the solutions that have existed for decades, right, is we actually do it very deep at the application level. So if you're an engineer, we just capture a ton of data that's really important to you about that error. So you don't really have to go back and say, well, you know, what actually happened or how do I reproduce this bug? And then more so we focus on what we describe the front end. So we focus on device-centric applications. So mobile apps, desktop apps, browser apps, which, you know, basically, these days, everything is a browser app. And it's actually become really important to have our kind of technology in those applications because, otherwise, those developers get no visibility, and they can't really solve any of those problems. But our solution does work everywhere as well. You know?

(Joel Beasley at 00:26:54) One of the things I thought was really cool when exploring the product was how it could connect like front-end application errors to back-end application errors. Two separate, like, you could do multiple applications, and then it had some intelligence with connecting errors. And another thing I liked about it was how, and I can't describe it like perfectly right now because I haven't been writing code full time in the past year because of the success of the podcast. But when looking at your error logging and then having to go another place to figure more stuff out, like, for example, like, I guess you would call like all that metadata about the sessions or all just all of the information. I guess I'm not describing it great, but if you go to the, what is it? Sentry.io? Yeah. If you go to Sentry.io and you look at the homepage, they've got like a great video, and then they've also got a couple of shout outs about different features. But it makes it very clear when you can see those pictures, they're like worth a million words. Right? But as a developer, I really believe what you're saying after exploring the product of you just constantly double down and make the product better and make it easier for the people to work so that they don't have to think. It's just like everything's right here, and it's simple, and it's also beautiful.

(David Cramer at 00:28:17) Yeah. It's been super important for us to, like, I don't know. I've always had to waste times in history where, you know, there is a bug on your website or something like that and, you know, if you're lucky, you have logs or something you can go look at. If you're unlucky or work at a real company, you don't get access to the logs. And even if you did, good luck finding the bug. Good luck having enough information in that log to actually understand what the bug was. Like, best-case scenario would be a little like bit of like the stack trace, which sort of tells you what happened, but that's it. And we always describe it. It's like sort of Sentry takes that, that stack trace. And it's always easiest to visualize it in terms of size, and it's like that stack trace is very small. It's like a kilobyte in size or something. And a lot of our payloads, like the information we collect to help make it easier to debug end up being like closer to a megabyte. So, you know, significantly larger, like, sometimes a thousand times larger. And that ends up being very significant because it is very much like, if we identify the bug, you always have enough information to be able to resolve the issue. And that's all it's for.

(David Cramer at 00:29:23) When we were developing our mission statement, because, you know, we just built something that was useful. We didn't crack it that way. We didn't do it intentionally, so we didn't have a mission statement or any of these other things that you'd have as a startup. Right? And when we're doing it, we just started working backwards from, like, what is the purpose of what we do?

(David Cramer at 00:29:41) And we just started with, like, okay, software only exists for efficiency, for productivity. It has no other value in the world. And so if that's true, then we are monitoring, which is where we cluster Sentry. And that means monitoring must exist for productivity. Otherwise, it has no value. And so that's how we reason about it. And so we're like, well, how do we make sure what we're doing translates to productivity?

(David Cramer at 00:29:57) And at the end of day, it's like, okay, you ship a new version of your application, it's broken. We want to tell you about it as quickly as possible so you can address it as quickly as possible. And we want to make sure you have the right information so you don't waste a bunch of time trying to address the problem. And that's it. It's pretty straightforward, but it is all about that efficiency.

(Joel Beasley at 00:30:23) Do you have any stats on time saving? Like, the distance between a bug happening and you understanding the root cause, without Sentry and then with Sentry?

(David Cramer at 00:30:37) It's tough. I could make up some stuff and tell you.

(Joel Beasley at 00:30:40) It won't be possible. No.

(David Cramer at 00:30:42) But generally, I've had stuff where I've spent a week debugging a cryptic error that was just so challenging to understand because of the way systems work. And generally speaking, Sentry—I don't know that Sentry's ever forced me to go down that debugging path ever since it's existed. I've always had enough information right then and there to be able to solve it. That might take me a little while to solve it, but I know what the problem is at that point. And even knowing what the problem is is often a challenge.

(Joel Beasley at 00:31:10) To make it so it's not something we're just making up, right? You could take a company, a customer, and take their resolution times, maybe in some third party ticketing system or something. Take their resolution times before Sentry, and then implement Sentry, and then take their resolution times after Sentry, and use that as a case study. Because those business people, they will buy time. Right?

(David Cramer at 00:31:31) Because that's—

(Joel Beasley at 00:31:32) What we're trying to do. And that was one thing that kicked up in the back of my—now I'm not a sales wiz, but from the past two years of trying to make money with sales teams—that kicked up in the back of my mind is, like, oh, this is the one thing. I see that you guys solve stuff faster and you've got the bugs there, but where's that? You could totally do that with a customer. And even your existing customers, you could go to them and say, go look back in your history and tell us the time of resolution now versus then, because that's a huge selling point.

(David Cramer at 00:32:02) Yeah. It's been interesting because we did a little bit of this back in the day. The true ROI is a little bit complicated. What we found is a lot of, at least our customers at the time, they just didn't care. Because the interesting thing about our business is because it is so bottoms up driven, and it's also this industry wide trend where, you know, it used to be IT teams buy products. You had to sell it to the VP. You had to get their buy-in, and then frankly, they just handed it down to the entire company or something like that. Right?

(David Cramer at 00:32:23) These days, it's like, oh, you want to buy something? Does it fit on my credit card budget? Cool. Go buy it. I don't care. As a developer. You know? And that's a newer thing. For whatever reason, it didn't used to exist where developers would get tools or budgets. It was only for IT software or production, like, servers or something like that.

(David Cramer at 00:32:35) And so a lot of ours has been like, well, the developer implicitly gets the value. They don't have to make a case for it, frankly, because we charge a fair price. So it's not like they're having to go in and be like, hey, this arbitrary thing is going to cost us $10 million. I swear to you it's valuable. No. It's just like we charge a fair price like a utility to some degree, and developers have been very much the champions of bringing it into the orgs.

(David Cramer at 00:33:06) And I think the few times where it's been a gray area and the price tag's been significant, the buyers have still understood the value. Sentry is—we do very well and we're, you know, market share wise, we're very, very large. But we haven't had to deal a lot with traditional thinkers. Like, I don't have a good example here. I talked to this guy who ran IT for a bank, and it's a bank you wouldn't have heard of.

(David Cramer at 00:33:37) And that's that traditional thinker part of it. It's not like a Wells Fargo or a Capital One. They're actually fairly progressive thinking. But it's like a bank where they're like, we're never going to use cloud software as long as I can. That kind of traditional thinking. Right? And, you know, they're not our customers, and they won't be our customers until they change their opinions about those kinds of things. Right?

(David Cramer at 00:33:50) And so because of that, we've been fortunate that we haven't had to do a lot of that sort of, you know, more formal or traditional marketing and sales. We're trying to do more of it now, of course, but, you know, it is an interesting benefit, but it's also a challenge because we do actually need to develop that muscle and be good at those things as well. We just have been fortunate we haven't needed it.

(David Cramer at 00:34:18) We just have been fortunate we haven't needed it.

(Joel Beasley at 00:34:20) Yeah. I mean, the thing that excites me about your company is I see you're growing. You guys, you know, couple hundred, close to couple hundred people right now. Right? That's around the area. But, I mean, I see you guys at thousands of people. And when you get to those thousands of people, I mean, selling value to non-technical people, obviously. And there's entire markets. Like, one market that probably is one that you guys won't approach for a while is maybe the government sector market because they seem to be so closed.

(Joel Beasley at 00:34:50) Everyone always has to make some physical piece. You have to find some excuse to make some rack they can install somewhere so that they can have it on-site. But, yeah. I mean, I'm really careful about companies that I hype up or that I'm pumped up about. And after, you know, exploring Sentry, I'm just like, I'm all in, and I think it's fantastic what you guys are doing.

(David Cramer at 00:35:13) Yeah. It's exciting. We're very much here for the long haul. When we raised money—so it was interesting. When we raised money five years ago, the world actually looked very different. People didn't quite understand developer tools still. They didn't understand open source or any of this stuff. It changed very rapidly, I would say.

(David Cramer at 00:35:24) And I remember these conversations with investors for our seed, and we had revenue. We were profitable when we raised our first round of funding. And I remember one of the questions somebody asked me. They're like, why would anybody pay you? And I'm like, they already pay us. What are you talking about? They're just like, no, but why would they pay you a lot of money? It's like open source. Why would they just use it for free?

(David Cramer at 00:35:50) And I'm like, who cares? We'll figure it out. That was kind of my answer. I'm like, don't worry about that. And I'm like, frankly, they're either going to pay us or they're not going to pay anybody else. That's all you need to worry about. And that was our methodology behind it. It was, like, create a great product, make sure it's accessible. And the open source thing was just another way to be competitive.

(David Cramer at 00:36:07) Even though we didn't build open source just for business reason or marketing reason, but that was very much our model going into it. And it's always been, like, let's build something great that is for every developer. And so, kind of from that seed stage, it was always, like, you know, to the moon, more so than, like, let's find it out. And so I think we have a long ways to go, but it's been cool to see what we've gotten to so far.

(Joel Beasley at 00:36:36) Yeah. I mean, as a developer, you know, I'll play with the open source version. But if I'm going to do something in my business, I want that premium support. I want it hosted. I want the versions, everything already being patched quickly. I want the commercial version of it, and they're usually reasonably priced. So it's not like you're spending, you know, hundreds of thousands of dollars to get this. It's usually just a, you know, couple hundred dollars a month charge or some premium version of some SaaS software. So it's not a big deal. Yeah.

(David Cramer at 00:37:04) Yeah. And I always think about it—I know a lot of organizations just like building things, and that's fine. But for the most part, we as a business, we prefer to not build stuff that's not our business. Right? And that's most people. So it's like, we just trust that people want to pay for the service instead of trying to become an expert on running it themselves.

(Joel Beasley at 00:37:21) Absolutely. Do you think AI is, or specifically machine learning, is going to play a larger role at Sentry?

(David Cramer at 00:37:29) We'll see. I'm a skeptic around a lot of this because what I would tell you is Sentry has a lot of AI. And I equate it—the best quote I've ever seen about this is if you take any sort of writing about AI and you replace the word AI with some if statements, it will still be very, very accurate. Because a lot of this AI and machine learning is actually just very simple models. And Sentry has a ton of models that humans have built.

(David Cramer at 00:37:54) They don't train themselves over time. They don't get better until we make them better. But they do create the same kinds of output that you would think of from an abstract way as AI. Right? And I think the area where that might change over time is, like, the more data you have, the harder it is to sift through it. And then I think even trickier is data becomes a little bit more unique to you.

(David Cramer at 00:38:13) And I think when those two problems kind of exist together, then you need something that's a little bit more powerful and that can be trained. And to me, until it has active training, it's not any true version of machine learning at the end of the day, or any AI. And I just think so many systems don't have that. They just run off of static models and Sentry's full of static models. But to date, we haven't needed anything like that. And I—

(Joel Beasley at 00:38:44) Well, I mean, I'd—

(David Cramer at 00:38:44) I'd say just—

(Joel Beasley at 00:38:46) You bring up a good point about static models. Sorry to jump in, but I want to bring this up because I struggle with this a lot as an engineer at first when people started saying AI because I'm like, that's a query. You know, it's a complicated query. They're selling AI when they raise money or when they're doing press releases, but that's not AI. You know?

(Joel Beasley at 00:39:01) And then I started asking myself deeper questions, like, you know, am I just being a curmudgeon? Like, what is AI? Like, how do you define it? And I've actually followed, like, some, I think, Lex Fridman and some other, you know, influencer type people that are really on the cutting edge, you know, writing papers and things about AI. And, I mean, I guess it's, like, its ability to reason off of a dynamically changing dataset means it's AI.

(Joel Beasley at 00:39:26) So that means I could write some query, like, a pretty simple query that reasons off of some dynamic datasets, and that's technically AI. But it's like a really weird line. Like, the more you know, the harder it is to define AI, and it's just fascinating how the world sees it versus how the practitioner sees it.

(David Cramer at 00:39:47) Yeah. Yeah. It certainly is. I think, to me, the challenge has always been a lot of what we're doing these days is not new. And I, you know, I'm also a little bit of a curmudgeon in the sense of, like, I hate bad marketing. I hate when people are selling me, like, uh, just frankly lies at the end of the day.

(David Cramer at 00:40:04) And so, you know, we have a lot of random marketing versions of this in our category and things like AIOps. I'm just like, what does that even mean? Like, it's—

(Joel Beasley at 00:40:19) Everything is not—

(David Cramer at 00:40:20) Product you had. Yeah. But, yeah, I don't know. We'll see. I think there is a lot of room for innovation in there. I just think practical uses for what I'm going to stick with is, like, real AI just are very rare, and nobody really achieves them. Obviously, like, what OpenAI and stuff is doing is, like, that's genuine, but those are not real products that people are consuming or getting value out of at the end of the day. They're still really interesting prototypes, but that's not what any of these companies are building. So—

(Joel Beasley at 00:40:49) No. I think some of the most interesting AI is the stuff that can take stories and then make new stories. Like, for me, that is the area where it gets magical because that reaches beyond my current education level. And I've done my Saturday night binge of reading the AI articles and reading papers and things like that about what they're actually doing with the corpus of text and things like that and all of that understanding.

(Joel Beasley at 00:41:23) But it's just such a fascinating thing when you understand how it's looking at the individual letters and looking at the words and figuring out the patterns, and then it can actually spit something out that's semi-cohesive. And it's like, that to me, whatever the underlying mechanics of this universe that are happening for that to occur is quite beautiful. It's really fascinating.

(David Cramer at 00:41:48) I do agree. And I think, you know, I don't remember what the quote is, but there's something—and sort of the test of AI is if it can convince you it's a human. Right? And I think that content is one of the most successful ways to sort of see, you know, a glimpse of that. Because, like, yeah, if it can beat you at chess, sure, but it's not doing anything else.

(David Cramer at 00:42:08) Whereas if it can hold a conversation with you, which if it can generate output based on a bunch of inputs, it could probably, you know, you can extrapolate and kind of reason that it might be able to hold a conversation with you soon. And then it feels like it's real. Right? Then it feels like what we as individuals, not software engineers or anything like that, might think of AI as. And that's sort of what I look for in it at the end of the day.

(Joel Beasley at 00:42:31) Yeah. Dude, learning to live is an interesting thing. Being able to find meaning and routines that work for you. What type of routines are you in right now?

(David Cramer at 00:42:46) Wake up. Go to the office slash the other room. Spend all day in the office. Go home slash the other room. I don't know. That's been the worst. Like, you know, we've started breaking out of this shelter in place for better or worse, and I just couldn't handle the monotony of, you know, kind of stay in your home as much as possible.

(David Cramer at 00:42:58) Like, we went to the lake two weeks ago. It had opened, but it wasn't too busy, which was good. Like, it didn't look like any of the photos of beaches you see on the news or anything, but everybody was socially distant. And that was so refreshing because the stay at home routine was just—it was killing me.

(Joel Beasley at 00:43:27) Yeah. It's amazing how much, like, just being around other people, even though you're not—you don't know them or you're not directly interacting with them—just having other people around impacts your mood and your sanity. It's unbelievable.

(David Cramer at 00:43:40) Yeah, yeah. I don't know. I was always the guy. I love the office, and to some degree that's cascaded in our company. It's very much an office company, and I truly believe it's much more productive. Maybe that's not for everybody, but for the majority, it's much more productive. And the stimulation you get out of conversation with your peers, the bonds you get even outside of work out of your peers, I think all that's super important. And I definitely really miss that. And that used to be my routine. It's just like up, commute, just start the day, really get stuff done, feel good about that, you know, leave, wind down with friends or at home. And it is very simple routine and consistent, but it worked really well. And these days, it's been tricky to find something that feels good, you know, that doesn't feel like you're wasting away to some degree. But, you know, hopefully this doesn't go on too much longer.

(Joel Beasley at 00:44:32) Oh yeah. There's this good YouTube video with Simon Sinek. Have you followed him at all?

(David Cramer at 00:44:39) Mm-mm.

(Joel Beasley at 00:44:40) Oh man, he is great. You're gonna love this guy if you check him out. But he talks about this concept of the different type of neurological reward systems, like dopamine hits and things like that, and how our body rewards us and what triggers those things, and how you can sort of set your routines up. And I personally experimented with this for several months about exploring when I'm planning my time, what type of physiological reward I'm going to get. Like, do I get a physical one for that feeling you get after you do a good run? That's a bonus bump up. You get the feeling of checking something off or completing a task. That's a different type. And so you've got—I'm butchering it—but you've got these four or so types of neurological or physiological responses, and you can, when you have those in your mind, you can look at how you've planned your day so that you're feeling really good.

(David Cramer at 00:45:36) Yeah, yeah, I definitely get that. I tried to—I bought a house last year and I've tried to become a little bit more handy. I've not succeeded, but I've tried to DIY some things or fix some things or improve some things, you know. And it's very unsatisfying when I fail to complete it, but a few times I've actually gone through the project all the way so far. I've been really good, and that's been a good way to spend time. But yeah.

(Joel Beasley at 00:46:02) Have you gotten the Nest yet?

(David Cramer at 00:46:04) No. But I did—I actually wrote a blog about this—but I did part of it. I basically wired up my home to do a bunch of things with smart devices per se, but it's not the mainstream stuff. So it's basic level stuff. But for example, when the washing machine turns off, it will announce over some in-wall speakers that the washing machine has finished. And so it's actually semi-useful things like that. Or I built something that would just automatically activate the alarm in certain scenarios just because it's really easy to forget that, and often the alarm software doesn't have routines built in. So I've tried to build things that make my life, I guess, more efficient, and that's been kind of fun. But it's a lot of just software stuff more than it is anything else. You know, go to the thing you know.

(Joel Beasley at 00:46:50) How did you do the washing machine one?

(David Cramer at 00:46:53) So there's a smart outlet from this company called Sonoff, which if you Google about this, you'll see endless versions of it. And what it does is when the washing machine turns on, it records the power is now on because the usage goes up. And then so when it's been on for more than a couple minutes or whatever, it just says, "Okay, washing machine's on." It knows the state. And then when the power usage drops to zero, it's like, "Okay, washing machine is off." It's very dumb, but it works perfectly.

(Joel Beasley at 00:47:21) That's amazing.

(David Cramer at 00:47:21) Yeah, it's actually super, super useful.

(Joel Beasley at 00:47:25) Well, because what did it do before? Because mine goes, you know, when it's done.

(David Cramer at 00:47:30) Yeah, so it does have a buzzer, but it's down in the garage so you can't really hear it.

(Joel Beasley at 00:47:34) Oh, oh, so you have a speaker. Yeah, I get it. Yeah.

(David Cramer at 00:47:38) And then the other cool thing has been, it turns out if you just send yourself a lot of reminders for things, that's actually really useful. So a lot of these smart home per se initiatives I've gone through have been like, send myself a notification—and my girlfriend who lives with me—send a notification to us on Telegram. And so we actually have this house thread on Telegram that the system just sends notifications to about useful events. Like if we leave the house and we don't turn on the alarm, for example, it reminds us. It's like, "Hey, the alarm isn't on and you don't seem to be at home. Do you want to turn it on?" So it's little things like that, which is actually very convenient.

(Joel Beasley at 00:48:13) Just program something in there where your house sends you a telegram that it misses you if you've been gone for more than 24 hours.

(David Cramer at 00:48:22) All the Easter eggs.

(Joel Beasley at 00:48:23) Raise some money and be like, "It's AI."

(David Cramer at 00:48:26) Yeah.

(Joel Beasley at 00:48:28) Dude, this is good. I like you, David. You're a really cool guy.

(David Cramer at 00:48:32) That's a good conversation.

(Joel Beasley at 00:48:34) Yeah. So as we start to wrap up, people can go to sentry.io, and you've got a free version of the product you can start with?

(David Cramer at 00:48:44) Yeah, so you can start free. We generally say free is if you're a hobbyist, if you have a side project or something like that. If you're open source or nonprofit, we often give you it totally free. But pricing is super fair. It starts at $30 a month. If you're a business, it's not a lot of money. You know, it goes up based on how much data you send. So the more errors you have, the more you're gonna pay us. It's good encouragement. But yeah, easy to get started. Usually up and running in five minutes.

(Joel Beasley at 00:49:13) Good domain name, easy to remember. Sentry.io. David, absolute pleasure, my friend. When this whole social distancing thing's over, maybe next year when I go to San Francisco—I go to San Francisco four or five times a year. I'm pretty familiar with it out there. I'll let you know. Maybe your office will be open. We could say hello.

(David Cramer at 00:49:28) For sure. Yeah, and thanks for having me on the podcast.

(Joel Beasley at 00:49:30) Dude, anytime. Cool. Alright, my friend, have a great day.

(David Cramer at 00:49:34) Alright, you too. See you. Bye.