The Impact Time Series Data Has on Companies with Mike Freedman, co-founder and CTO of Timescale

(Intro Narrator at 00:00:03) Hello, my friends. Today, we're talking to Mike, co-founder and CTO of Timescale. And we discuss the impact of time series data, how Mike's experience as a professor gives him insight as a CTO, and how T-shaped people are getting ahead professionally by leaning into their strengths. All of this right here, right now on the Modern CTO Podcast.

(Joel Beasley at 00:00:32) Here we go. This is the Modern CTO Podcast. So dude, why do you need time series specific databases?

(Mike Freedman at 00:00:46) You know, it was interesting. When my co-founder and I started, we were actually looking at this broad problem of all these devices were coming online, all this operational data. You know, I think the last fifty years were what they call digitizing the back office. And now everything's getting digitized. It's houses, it's people, it's trucks, it's manufacturing lines.

(Mike Freedman at 00:01:11) And so you think that often the value from this is actually the data coming off. It's not that they're connected. It's that they're connected and sensing, and we could do things with them. And we want to build then applications. Developers and teams want to build applications that use this data.

(Mike Freedman at 00:01:27) And what do you need for a core part of any type of data-rich application is a database. And so when we actually started this, we were more specifically looking at IoT. But when we ended up building initially our database for IoT, we came to realize that actually the database that we built for the data-driven applications for IoT is much broader. And, you know, at Timescale, we now have this database that can solve problems for not just IoT, but for observability and for crypto and for fintech and for marketing and for logistics. Basically, we like to say is that all data is actually time series data.

(Mike Freedman at 00:02:07) If you think about it, captured at the finest granularity, it's about the rate of change. And, you know, great data products need a great database. And so that's what we've built here at Timescale.

(Joel Beasley at 00:02:16) So my background, software engineer for seventeen years. Then I started this podcast, so I haven't been actively writing code for about three years. But would it be a separate database just for time series specific data, or is it all one database? Explain that to me.

(Mike Freedman at 00:02:30) More specifically, we're actually a relational database that scales for time series data. And we're actually built on top of Postgres, which is, you know, this amazing open source, community-driven product. And we like to say that we supercharge Postgres for time series data because what we find is actually it's not just the sensor data, right? It's not just the metrics or the events.

(Mike Freedman at 00:02:52) You actually often have other data, other relational data, other business data around it. And so often what we find is that people want to store inside a single database, or possibly on our cloud across multiple databases that talk together, both the time series data that they're collecting—the metrics, the sensor data, the events—as well as all the information around it that really makes it valuable for the applications they're building.

(Joel Beasley at 00:03:19) And you said it's open source. People can go sign up for it?

(Mike Freedman at 00:03:22) Yep. You could go on GitHub. You could download and play with it. We also, from a business perspective, Timescale offers Timescale Cloud. So we offer a fully managed and really a whole product experience around what developers—how developers want to build modern applications into managed services in the cloud. But, yeah, people could also just go to GitHub and download it.

(Joel Beasley at 00:03:41) I like that because a lot of the time, what I would do in the past when working on projects, you know, you get a project, then you have to figure out the toolset that you're going to use to accomplish the goal. And so you've got to learn a bunch of new technologies and tools and whatnot if you're an agency-type company. And then I would really enjoy it when I could find one of the open source solutions that I could download, play with, and then I just go use the paid hosted solution simply because I don't want to manage updates. And there's a number of small items that specifically were important to me. Whenever I see companies come across with that business model, I'm always a big fan.

(Mike Freedman at 00:04:15) Yeah. You know, it's interesting. Two things. First of all, you say you could use open source because you've got to get all the tooling. I think at Timescale, it was even better because we build on Postgres. It's like all the Postgres tooling works with TimescaleDB. So any application framework, any ORM, any visualization tool that speaks basically SQL, that speaks to a Postgres database, just works with Timescale out of the box. And that's been a real—one of the biggest ecosystems out there. But to your point, you know, it's interesting from, again, from a business perspective. We, from a commercial perspective, we really focus on the cloud. So we don't try to then sell you an on-premise thing. All of our database is completely free.

(Mike Freedman at 00:04:55) We don't have a proprietary version on-premise that we try to kind of upsell to people. And so we think enough of a modern developer wants a full cloud experience. It's not just about hosting. It's about the entire experience that we build around that and make you more productive as a developer and allow you to build applications better, cheaper, faster. But for those, you know, good segment of the market that wants to manage themselves, wants to do dev themselves, want to stick things in their CI/CD pipeline.

(Mike Freedman at 00:05:25) You know, that's just, you know, all of our free open source version.

(Joel Beasley at 00:05:31) Tell me about how can people try the product?

(Mike Freedman at 00:05:34) One of the easiest ways, we actually offer a kind of a free, unlimited, no credit card required, thirty-day trial for Timescale Cloud. And if, you know, you just go to timescale.com, click on a big button at the top, you could be taken immediately into the cloud experience. Within literally thirty seconds, you'll have a database up and running. You can immediately connect to it and you're kind of on your way. And, you know, we found that has enabled people to become successful really quickly and certainly see all the value with their time series data and with TimescaleDB.

(Joel Beasley at 00:06:05) What's an example of some of the benefits that sit on top of Postgres?

(Mike Freedman at 00:06:10) Sure. So there's a number of different things and they're generally speaking big buckets. You know, we effectively—when you think about time series data, one of the things is there's just a lot of it. So you might think of a normal database, a normal table getting big in the millions, tens of millions of rows of data. We have people that in what we call a hypertable—it's basically it looks like a Postgres table and you could treat it as such, but we do all this magic under the covers.

(Mike Freedman at 00:06:38) We have people that'll store a trillion rows of data in what looks like one table or, you know, terabytes of data in one table. And so we have a lot of things that focus on scalability, you know, make it really scale to those—you know, I don't know. You said you were a developer for, was it fifteen years, seventeen years?

(Joel Beasley at 00:06:57) Yeah. Seventeen years.

(Mike Freedman at 00:06:58) Seventeen years. Like, you know, most developers don't deal with databases of hundreds of billions of rows. Certainly not frequently, but that's common when you have a constant stream of data. And so how do we make it really scalable without any duct tape and glue like a lot of solutions are? How do we just make it elegant?

(Mike Freedman at 00:07:16) How do we make it much more performant for the type of queries, the analytical queries that people want to do? How do we build—we build kind of automated, the technical term is incrementally materialized views that are constantly kept up to date. But, you know, you could generate summaries. You know, you have this underlying raw data, and then your business problem is I want to build real-time dashboards, or I want to give this view to my customers. And this is one line of description and then we keep that all up to date, and it's even kept up to date if you backfill data because sometimes data arrives late, and we have all this technology under the covers that keeps it up to date.

(Mike Freedman at 00:07:53) We have really advanced compression that's 95 to 98% compression and so, you know, really massive cost savings. And then a lot of analytical capabilities that are kind of unique to doing the type of queries that people often have in terms of sketching functions and downsampling and approximation functions and special algorithms to make it visually interesting to support visualization. So it kind of all comes built in, but you speak SQL, it looks like a normal database, and you're kind of off to the races really quickly.

(Joel Beasley at 00:08:26) So you can generate visualizations for this data?

(Mike Freedman at 00:08:28) Yeah. We don't build a visualization tool ourselves. But because we just speak SQL, virtually any off-the-shelf visualization tools work, you know, Grafana, Power BI, Superset. A lot of people use us both for internal visualizations, and people use Timescale also for customer-facing. We are often the serving layer of their applications.

(Mike Freedman at 00:08:50) And so sometimes they use those off-the-shelf for internal, while they often use then—they build custom visualization and dashboards, whatever, that's branded for their own customers. And because we just look like SQL and Postgres, it just works out of the box.

(Joel Beasley at 00:09:04) Nice. So main thing is expandability for the volume of time series data that you'll have.

(Mike Freedman at 00:09:10) Scale and performance are a huge thing while still being, you know, very cost effective from compression and some new things, additional things that we're building in the cloud. And really worry-free. You know, you sit on this bedrock of twenty-five-year-old technology of Postgres that has been, you know, battle-tested through the decades.

(Joel Beasley at 00:09:29) Nice. Yeah. I've had some conversations about that. I personally have only experienced it a little bit when I was building a financial software tool that did portfolio predictions for different events that would happen. And, you know, they would have to figure out what would happen across these twelve scenarios.

(Joel Beasley at 00:09:48) And then another part of it was, here's all the attributes of anyone's portfolio. Now figure out, all taxes considered and everything, figure out the best way to withdraw this over time. So we'd have to run these simulations constantly to figure out what's the most efficient way to run this over time. So I got to learn a little bit about it there, but I've learned a lot talking with some of my friends that work at marketing companies.

(Joel Beasley at 00:10:13) where they're tracking clicks and things like that and huge amounts of data. And then it's financial software and then marketing are two of the areas where I've seen it myself a little bit.

(Mike Freedman at 00:10:22) Yeah. I said before, we started working on devices and IoT. And then when you look around, we like to say internally that, you know, all data is time series data. And it's because anything that you record, you can either record the latest state or you could record the change over time. And it's kind of like, what is more valuable to you, a photo or a video?

(Mike Freedman at 00:10:41) What gives you more information? And to be able to have, you know, a data technology and database that makes that efficient and enables you, it just unlocks a lot of value to companies. You bring up a couple examples, you know, marketing tech, fintech. We actually have a lot of users in all those areas. And whether or not it's traditional stuff like banking, whether or not it's traditional portfolio analysis, backtesting, you know, what a lot of hedge funds do, and then more modern fintech companies.

(Mike Freedman at 00:11:10) And we all see this in crypto and in Web3 and in NFT. A lot of those core technologies really apply to many of these things. You know, if you've ever logged into any type of, whether or not it's crypto or traditional stocks, you'll see these candlesticks where it tracks the highs and lows. And often that's, you know, one minute, fifteen minute, one hour, one day. We can look at different granularities.

(Mike Freedman at 00:11:35) Well, I was talking before about these incremental views, and that works perfectly for this because you just say, let's actually create an automated rollup for one minute, thirty minutes, six hours, and the database already handles that. And then the serving layer to your customer is literally just these rollups that the database is constantly keeping up to date against the raw data. And you don't have to actually build these as ten different pipelines. You know, we really kind of simplify the backend infrastructure that Timescale can provide for your own customers.

(Joel Beasley at 00:12:06) What's the pain point? Because there's going to be a technical pain point, and then that's most likely going to bubble up into some sort of business or customer-facing pain point. What are they experiencing before they find you as the solution?

(Mike Freedman at 00:12:17) Yeah. So the biggest thing that we found, again, is because the existing market of traditional relational databases is just so massive. You know, the millions of companies using technologies like Postgres or MySQL or sometimes Mongo, though not a relational database, that were not designed for time series data, either the scale, performance, or the type of analysis you want to do. But they're what every developer knows. So often what happens is they pull them off the shelf, and what they soon learn is this isn't scaling for us. It's not satisfying what we need.

(Mike Freedman at 00:12:52) So a huge number of our users are really people who started on vanilla Postgres or MySQL or Mongo. They started building their applications. They realized it doesn't scale with their needs, performance, and all these things. And then they actually moved to Timescale. But, again, because we're at a core relational database, it's not like they have to completely, you know, change this thing where they're now having a ten-microservice thing speaking MapReduce and little Erlang functions and streaming things.

(Mike Freedman at 00:13:25) And so, yeah, you can do things a lot of ways, but we're technology that is operationally simple and yet allows you to scale and not really worry about it. So that's the biggest thing that we find. Sometimes, there's also a lot of the other time series databases that are much more niche solutions. They kind of typically came out of the observability space where they're really just tracking things like CPU, and so they're much more limited. And so what we see for people who might have said, well, I have a time series problem.

(Mike Freedman at 00:13:58) I need a time series database. Well, if you go back to what I said before, it's not just about that metric data. It's often about the other data around it. So where a lot of other people start is they say, well, I need to deploy a solution with a niche time series database plus a database like Postgres or MySQL. Then I've got to write application code to join it.

(Mike Freedman at 00:14:20) And when they come to Timescale, they could actually put it all inside the database, and it's much more—the pain point, it both solves their scale problem, and it's operationally much more simple, effective, cost-effective, you know, performance, et cetera.

(Joel Beasley at 00:14:33) Yeah. That's a brilliant observation as far as the entrepreneurial aspect of you being the co-founder. You see these people putting these things together, and you put them together, and you save them a lot of time.

(Mike Freedman at 00:14:45) Yeah. And the thing what I talked about in the beginning is we initially had built TimescaleDB because we had this problem ourselves. You know, we were initially looking to solve this data problem around IoT, and we weren't starting to say, like, hey, we need a new time series database. We thought, well, this new wave of computing, how do we build a data platform for that wave of computing?

(Mike Freedman at 00:15:07) And in kind of looking at the market, we found this niche solution for time series databases, but it couldn't satisfy our need. It wasn't performant enough. It didn't allow us to store all the metadata that we wanted about IoT devices that you need. You want to ask questions not just about what metrics, but what device? Where is this device located?

(Mike Freedman at 00:15:26) How do I slice and dice it by different attributes? So we ended up building TimescaleDB for our own needs for that IoT data problem, and then soon realized, and what we heard from a lot of users, is how that problem was basically everywhere. It wasn't just about IoT. It was about music analytics and trucking and personal health and connected health devices, and basically everywhere.

(Joel Beasley at 00:15:56) Yeah, now that you're bringing it up, I'm thinking it's a lot of places. I'm gonna have a hard—thanks a lot, bro. I'm gonna have a hard time of not thinking about things in time series data now. See my Ring doorbell and all those things.

(Mike Freedman at 00:16:07) You know, like everything else during the pandemic, we got our Peloton. And if you just think about Peloton, there's the internal device metrics that the company is collecting—time series data. There's your leaderboard—time series data. There's your performance over time—time series data. There's the video analytics—time series data.

(Mike Freedman at 00:16:27) And it really is just everywhere that you want to go, and you need a solution to build for that.

(Joel Beasley at 00:16:32) Yeah, that's really, really cool. Now the company before that, sort of wherever you were that inspired this and found this problem, does that IoT company still exist? Is it something separate?

(Mike Freedman at 00:16:42) We were working on that for, you know, it was only about a year or so. And we had some moderate success. We had a number of, kind of, a hundred thousand devices sending us data. But really, Timescale was born out of what we had built for that entity and really just saw a much broader opportunity at Timescale.

(Joel Beasley at 00:17:03) And so how long has Timescale been around for now?

(Mike Freedman at 00:17:05) We launched in 2017, so it's been around five years. We were building the database for a bit before that, but we were initially IoT. And then the minute we launched the database, just the amount of interest and the velocity of the adoption—it was really clear that this is what people really needed, developers really needed.

(Joel Beasley at 00:17:30) And was this your second company that you founded?

(Mike Freedman at 00:17:35) Yeah. So, my background is I'm also an academic. I'm a professor at Princeton. At grad school, I did a lot of research in distributed systems and content delivery networks. I ran a CDN, an open free CDN, for about five or ten years with lots of free users.

(Mike Freedman at 00:17:52) And in grad school, I did some work around what became kind of IP geolocation, and it was a small company that was acquired. I did other work that led to—if any of you have heard of software-defined networking and OpenFlow—I was part of the team at Stanford that developed some of that. And there were a bunch of startups, a lot of industry built around that, but I kind of moved and started more of an academic career at Princeton. But my heart's really been in building product and building things that people use. And even the products I did, I wasn't satisfied with an academic paper.

(Mike Freedman at 00:18:25) I really wanted to get it in people's hands and kind of change industry. And so, in different parts of your career, you get kind of—you want an interest in doing a new thing. So a number of years ago, I said, you know, five or six years ago, I reconnected with my cofounder, Ajay Kulkarni. We've actually known each other for twenty-five years. We met the first week of college.

(Mike Freedman at 00:18:48) Both of us had started businesses before. He, at the time, was at Microsoft from a bunch of acquisitions from his last company, and this was really a great opportunity to tackle this new big data problem we saw.

(Joel Beasley at 00:18:59) That's pretty cool. So you're still actively teaching?

(Mike Freedman at 00:19:02) Yeah. Although, you know, my research output probably isn't what it used to be. But I think in the end, the job of all of us is to figure out how we could build great things and have an impact in the industry and the world. And I think at Timescale, we like to say we really kind of help developers build the next wave of computing. And so I think that's a great mission for both the company.

(Mike Freedman at 00:19:22) I think it really excites our team as well as certainly myself personally.

(Joel Beasley at 00:19:26) You actually have a fairly interesting positioning because your company has a hundred-plus people, and so you get to watch it grow, and you get to watch the generations of people grow within your company. Outside of that culture, you're also seeing, like, next generation in the classroom. So you've got, like, two different views where most people only will have, like, one, like, within their company.

(Mike Freedman at 00:19:48) Yeah, it's interesting. I mean, I was also doing one thing for about fifteen or so years. And it's always fun in different parts of your career to take on new challenges. Being a professor at a research university, you know, while certainly I do teach undergraduates, a major portion of that is also running a research group, which is kind of almost like the closest thing is an old apprentice model where you only have a few people and you work deeply.

(Mike Freedman at 00:20:14) And I think, you know, if you want to do right for your students, which I think is really important, I think what makes a person successful in an academic career and what makes a team successful are very different things. Academia is a lot about personal accomplishment, and so you want to set up your graduate students to really be the owners of the research they're doing, not just like one—at least in academic computer science—not like one person in a major project involving twenty people. You want to set them up to be the leaders in the project. But that also leads to a certain scale where you kind of work on, let's say, smaller things, where if you want to really have big impact, I think that's one of the really exciting things about building startups is that you could actually have a small team from the beginning and then grow and really kind of try to have industry-wide impact.

(Joel Beasley at 00:21:05) And I'm learning about you, right? So I was all excited about this, like, dual—you know, I do this podcast. I've been doing it for five years, right? And so I've done five hundred-plus of these interviews. And one of the things I personally like to do is figure out leadership stuff, right? Like, what's common between these leaders? And so for me to sort that appropriately amongst my experiences, I always try to figure out, like, what type of environment they're in just so I can sort of, like, help categorize the data.

(Joel Beasley at 00:21:33) So my question that I wanted to ask is, what is going on with the next generation? Is it different? Are you interacting with a lot of people, like, more than twenty people at your professor job?

(Mike Freedman at 00:21:46) Yeah. I mean, often classes I teach are a hundred-plus.

(Joel Beasley at 00:21:49) Okay. So what are you seeing in the next generation of leaders there?

(Mike Freedman at 00:21:54) Well, I mean, I would say two things. One is that a lot of the students are very passionate about entrepreneurship. And I think similar to myself, they're often optimists about the potential for technology to bring around good change in the world. And I think that's related, similar to what you see in a lot of tech companies and startups, is the rate of change, like how quickly technology can be launched. And given especially all the ecosystems of cloud computing and other ways and financing that we have to build and scale companies quickly, it has really excited, I think, the younger generation and the ability for them to have immediate impact as opposed to, like, I'm gonna go work at Microsoft or something and be one of 200,000 employees or whatnot.

(Mike Freedman at 00:22:47) The other thing I would say is I think that the rate of change of technology really surpasses what we teach in school. I think there's good reason why colleges and universities teach fundamentals, but it's interesting how fast all of the technologies and toolings around the web really surpasses what we teach, and the velocity of that change is so fast. So it's kind of interesting in that, like, I think it's still good that we teach fundamentals and, you know, at Princeton we get a lot of smart people. And if you put a bunch of smart kids in a room and teach them stuff, they generally do well. So we don't have that hard a job in that regard.

(Mike Freedman at 00:23:28) But there's a lot of learning that really goes on outside the classroom, I think, is really interesting. I think that's also why you've seen the growth of all of the boot camps and different types of alternate learning spaces where people don't need—it might sound strange to me as a university professor. Like, I think people could be wildly successful in this industry from very atypical backgrounds. And actually, that's one thing I'm proud of at Timescale. We have people from a lot of different backgrounds in terms of learning and education, and we're a global remote company.

(Mike Freedman at 00:23:59) And I think that type of diversity on many levels has really strengthened our company.

(Joel Beasley at 00:24:05) I like that. Yeah, I have a non-standard background because I didn't go to college. I did an early exit out of high school in eleventh grade, the same program that the homeschool kids would exit from, because I was building software online. And it was my career, and I convinced my assistant principal and my parents to let me exit school early so I could devote more time to building software systems.

(Joel Beasley at 00:24:29) And it worked.

(Mike Freedman at 00:24:31) I think that's true for a lot of us. Like, you know, now they teach computer science early on in high school and even before. And, like, I went to public school and we didn't have any of that. And I learned my programming by programming for random things on the Internet and the dial-up Internet. So I think it's interesting in that it's become most ubiquitous now.

(Mike Freedman at 00:24:54) You know, turning back to—people think of Princeton as a humanities and all that stuff. By far, the most popular major at Princeton is computer science. And it is the most popular classes, the most popular major. And so, you know, it's funny how ubiquitous it's taken over everything, you know, while I still think that there's really so much more growth we could have from people of all different types of backgrounds.

(Joel Beasley at 00:25:18) Oh, yeah. Absolutely. And people that have, you know, specific experience and that really care what they're doing. I mean, that's what I found as an entrepreneur, and you can tell me if you find it to be true too. But I definitely look for the people who are really curious, who are persistent, who have discipline.

(Joel Beasley at 00:25:38) I look for these sort of, you know, traits, and then I figure out, like, how far off their knowledge base is based off of what I need them to do. And that sort of helps determine if it's, like, if it's adjacent, if they're really close, I'm like, alright, let's hire them and then they can figure out how to get there. If it's too far away, then probably not so much. But when you're hiring people within Timescale, what are you looking for?

(Mike Freedman at 00:26:00) I think, you know, what I like to think about is how do we build strong teams, not necessarily only the individual. And you talk about looking to how individuals—what are their skills? What are their mindsets? You know, I trust my teams to interview for technical abilities and whatnot. Although we've thought a lot about how to build an interview process where we don't do LeetCode.

(Mike Freedman at 00:26:26) We don't do any of that stuff to make it seem like, how are you gonna do your day-to-day in our type of environment? But, you know, what's one thing that's interesting is, we build a database. We build highly technical cloud infrastructure. Our customer is the developer. So the great thing is that we're able to attract engineers who are kind of passionate about what we're building because they understand the product.

(Mike Freedman at 00:26:48) They're often users of the product. But I also think that one of the things I really focus on is even though we're a developer-centric company, we want to think of what we're building, you know, really customer and product first, not are we building this because it's cool technology or because it's cool engineering. And so what that means is when we're thinking about how to hire and how we build teams, I really look for this balance of people who are, you know, who are really passionate about the technology. We have a lot of people that say, I love databases. I want to work on databases internally.

(Mike Freedman at 00:27:19) Coupled with people who are very product-centric and very customer-centric engineers, I mean, who are thinking like, well, there's various trade-offs, but, like, what is the actual engineering? What is the customer? What is the user problem we're all solving in it? And how does this feature I'm building really map to the needs of my users? We sometimes talk about that internally as outside-in versus inside-out thinking.

(Mike Freedman at 00:27:45) You know, inside-out is we know our technology deeply. What do we need to do? But our users don't. So how do we try to put ourselves in the mindset of our users, particularly of the new users to the product, and think outside-in? How are they gonna experience this technology, this onboarding, this new user experience, and try to get everybody aligned around that?

(Mike Freedman at 00:28:06) So to your question, you know, I really look for how to build great teams. And, yeah, that team is made up of great individuals. But in thinking about teams, I often think about, you know, this is what the current team structure looks like. How do we look for individuals who are going to complement that with the different folks and different skill sets as opposed to just, you know, more of the same kind of core background?

(Joel Beasley at 00:28:31) And what are you learning right now as a leader?

(Mike Freedman at 00:28:33) You know, I think that we're kind of in that hyper-growth stage of Timescale. And, you know, in a short number of years, we've—both the number of users and customers, but also just in terms of people—you know, we've grown in about eighteen months from about forty to about 180 people. And, you know, that puts a lot of new skills as a founder and leader of what you need to do. I think I've always been a bit of a problem solver, and I think a lot of engineers are. A lot of—although I don't like the term—technologists are, certainly academics are, you know, try to solve problems.

(Mike Freedman at 00:29:09) And that really works well in small groups and with smaller problems. You know, when you scale, you can't be the problem solver for everything. You need to think about how do you actually more paint that longer-term vision and how do you create the structure where other people are able to do that. And yet realize that while you're in that fast-growing, hyper-scaling notion, you're bringing a lot of new people onto the team. Half the company has been there less than a year.

(Mike Freedman at 00:29:39) And so you need to do that balance between how do you give people that space to think about, constantly, how do I align them to where the company is going into the broader vision because many, most of the people are new. And so I think that's always the tension and the challenge of any startup in our position, and certainly that's something that I've been going through myself as a leader.

(Joel Beasley at 00:30:05) Tell me why you don't like the word "technologist."

(Mike Freedman at 00:30:08) Well, it's probably the same reason why people who have LinkedIn bios of "thought leader" and "futurist"—

(Joel Beasley at 00:30:18) I love it.

(Mike Freedman at 00:30:19) Maybe I'm gonna get all these angry tweets or angry LinkedIn DMs. Yeah, in the end, I like to work on problems, and I like to solve problems for people and build things that help change the industry.

(Joel Beasley at 00:30:35) Yeah. We have that in common, man. So, dude, tell me a little bit about what your family life is like, being a founder and starting this business over the past five years.

(Mike Freedman at 00:30:45) Actually, the pandemic was particularly challenging for us because my wife's also an ER doctor.

(Joel Beasley at 00:30:51) Oh, okay.

(Mike Freedman at 00:30:51) So not only did we go through everything with children, but obviously she was ground zero for a lot of things that was happening with the pandemic. But I think that — and Timescale today is a fully remote company — and I think that actually has been really a great ability for us to attract... I mean, lots of companies tried to go remote, but when you're committed to being fully remote, it allows you — you have to think carefully about how you want to structure your teams and structure communications and deal with — it's not like we're remote, everybody a hundred miles from one city. We're around the world in 26 countries and six continents. And when we deploy something that's Timescale Cloud, people trust their business. The great thing is we have follow-the-sun operations and support across the entire world, which has been reflected greatly in how we could make sure we build a reliable product for our users.

(Mike Freedman at 00:31:51) But it also has allowed the flexibility for people with their home life to — and especially because we're globally remote — it allows us to think about how people can structure their working hours around also things like family commitments that don't always run nine to five. And the fact that we are actually already globally distributed means that most people in the company, there's not one nine to five that the whole company operates on. So it's really allowed a level of flexibility that I think has been greatly rewarding both for me personally as well as for a lot of people in the company.

(Joel Beasley at 00:32:30) Here's the thing that I was thinking about when you were speaking. We had to intentionally decide we were going to be a remote company. We couldn't be on the fence. I couldn't have the office and maybe have the people in the city come into that one and then have them half remote. I don't see that working well at all.

(Joel Beasley at 00:32:45) The hybrid stuff is difficult at best to do. But we made a conscious decision and decided to go fully remote. And honestly, I loved it. I'm a big fan. We've been this way for at least two years now.

(Mike Freedman at 00:32:57) Yeah. We initially actually also had a headquarters in New York City, and we had an office. And maybe for a while, about half the company was physically located there. And as soon as the pandemic hit, we closed that office and really went fully remote. And we've also scaled a lot more since then. I think that's the way to do it. You have to be intentional. You have to go all in, and you could kind of build a company much more about that. A hybrid is the worst of both, if you will.

(Joel Beasley at 00:33:26) Yeah. Yeah. You get that second-class citizen thing that happens. And it's funny when you get to read about something and then you get to experience it, and you're like, oh, that's so true. Because at first when I read it, I'd read somebody talking about the experiment that they ran with that, and then that's why I learned that second-class citizen term.

(Joel Beasley at 00:33:44) And I was like, that's really interesting. But it's probably not that big. Like, if you have the right people and you have the right culture, I'm sure it wouldn't be that big of a — nope. It happens, and it's hard. And for me personally, remote was the right choice at that time, and it still is today.

(Mike Freedman at 00:34:01) Yep. Yeah.

(Joel Beasley at 00:34:01) You're a smart guy, man.

(Mike Freedman at 00:34:04) I don't know. I think when you build a great team, and I think that's one of the things I'm most proud of, is we've been able to really get a lot of people who are passionate about it. I think, you know, I talked a little bit before — I think one thing that's also very exciting is we like to say at Timescale that we're helping kind of build the next wave of computing. And I go kind of back to those applications that we talked about. It's not that we're trying to allow you to take your back-office Oracle application and migrate it to something else. Though there's nothing wrong with that and a lot of businesses need that. But I think when you look at the world, what's the type of things people are going to build over the next ten or twenty years, a lot of them are these modern applications, and we're kind of — I mentioned before — fundamentally technology optimists about how we could use technology to kind of change the world. And a lot of that we're doing that. We have companies that are deploying sensors to track air pollution in Delhi and earthquake detections in Chile.

(Mike Freedman at 00:35:10) We have people who are monitoring the manufacturing lines of some of the largest global manufacturers of EV batteries for cars. And so you look at all these applications, and I think it really gets the team excited about our ability to kind of build infrastructure to help enable all this type of change.

(Joel Beasley at 00:35:29) That's pretty cool. Yeah. That — oh, this is a great plug — if people are interested in learning more about working at Timescale, where would they go?

(Mike Freedman at 00:35:37) Timescale.com/careers. We have — where we continue to grow broadly across the whole company and we hire globally.

(Joel Beasley at 00:35:44) So you're hiring in multiple roles then?

(Mike Freedman at 00:35:46) Yes.

(Joel Beasley at 00:35:47) Cool. What are you personally excited about the next milestone for Timescale?

(Mike Freedman at 00:35:52) I think that the first couple years, we really focused on kind of building the core database. This was also an interesting change as we spent the first few years building the core technology of the database. We always thought that ultimately where the market is going is that people want to consume these technologies as managed services, as cloud experiences. But, you know, we weren't — there are companies who take all these open source and the company's job is just running existing open source things. We are both building the underlying technology and then also building the cloud infrastructure that's a core part of.

(Mike Freedman at 00:36:33) And so we actually have, interestingly, almost very two different teams, two different personas of both database internal engineers and kind of cloud engineers and all the product experience around that. And so we really launched our core product, our cloud product, about two years ago. But more broadly, if I fast forward five years, I don't think what people want is, hey, I just want a managed database in the cloud. We think about, like, what is the experience of a whole database cloud?

(Mike Freedman at 00:37:04) Because when you're an application developer, if you think of even a modern microservice architecture, you don't run a single database for all of your microservices for all of your needs. You often have lots of these individual databases, but often they have data that is useful to bring together. And so the great thing about Timescale Cloud is we kind of enable that. It's not just about how do I run TimescaleDB, but how do I enable the entire fleet of your databases that need to exist, and how do they share data all between them, and how do I think of all these cloud-native tools like general fast storage and object storage, and how do I make that cost-effective in performance? And so we make the user experience really well.

(Mike Freedman at 00:37:49) It's SQL. It looks like tables. You could do joins. Over the covers, we've done a lot of the heavy lifting for you that kind of couples together all these different technologies and really kind of exposes that unified user experience. And that frees you up to really focus on your problem, like, build the type of application that you want and, like, kind of us doing a lot of that underlying heavy lifting.

(Joel Beasley at 00:38:11) I've got a fun question for you. Is that okay?

(Mike Freedman at 00:38:13) Absolutely.

(Joel Beasley at 00:38:14) Hypothetical. You're driving to the ER to have some lunch with your wife, and you're at a stoplight. And up next to you, this guy pulls up, right? And the car next to you rolls his window down. It's Elon Musk. And he's like, hey, Mike. I want to show you something. I just built this new invention. Come check it out.

(Joel Beasley at 00:38:33) So you go to Elon Musk's house, super cool house, by the way, and he shows you that he has built a time machine. But you can go back to yourself and tell yourself one thing right when you were starting college. The first week when you met your co-founder, you can tell yourself one thing. What would it be?

(Mike Freedman at 00:38:52) So this is Elon Musk. So it has to be actually — bought when I first learned about Bitcoin in 2012 or whatever it was, or at 10 cents — it was buy Bitcoin. No. I actually used to do crypto. So it's funny. I worked on the first e-cash system at a company in the year 2000. So it was kind of crazy.

(Joel Beasley at 00:39:09) We'll let you off easy with buy Bitcoin.

(Mike Freedman at 00:39:12) I'll stick with my buy Bitcoin at 10 cents.

(Joel Beasley at 00:39:15) There you go. Now I made a note when you were talking earlier because I didn't want to interrupt you. You were making a good point. But you said this phrase I hadn't heard. I think you were talking about team organization or something. You said follow the sun. What is that?

(Mike Freedman at 00:39:28) Yeah. So the idea that the sun's around the world at all points. So typically, if you're running basically a full support model or if you build important infrastructure that people rely on, it really often means that you have shifts of people who are working at all hours. Because when something happens to go bad, you want somebody within five minutes to be able to step in and do it. And as organizations begin to scale, large support or large operational organizations often set up centers around the world.

(Mike Freedman at 00:40:01) And one of the reasons they do so is so that they could actually have people who are basically working on their local hours. But it basically is available 24/7. And as a fully remote company, we're able to basically provide that kind of much earlier than a traditional company who is kind of office-based. We have people in Asia and Australia, Europe, Africa, South America, North America, West Coast, East Coast. Because in the end, I talk about this a lot — these applications — and what we found at Timescale is our customers rely on us for their mission-critical applications, for their customer-facing applications. We've had people say when Timescale goes down, our manufacturing lines stop. When Timescale goes down, we're no longer trading in the market. Our support engineers get paged at night if we were ever to go down. And so because of that, we actually take a lot of — we take that seriously, and we think about how do we build the practices, the processes, the organizational structure where we kind of live up to the needs that our customers place on us.

(Joel Beasley at 00:41:10) That's smart too because if you have that set up, you don't have to necessarily wake up engineers out of the middle of the night, right? If you have engineers in all the different time zones. I've heard of that, but I'd never heard it called follow the sun. And whenever I hear something on this show, even if it's the stupidest thing and people are laughing, I ask people to explain it to me because if you explain it to me, Mike, you're a professor, I'm never going to forget it.

(Mike Freedman at 00:41:33) Right? And we're really proud of the team, both the engineering, but the operations and support team. Really proud of the quality of people and dedication that we have. And that's really been a great experience we've had here at Timescale.

(Joel Beasley at 00:41:45) So if there are people listening that aren't necessarily software engineers, you also will hire customer service people that aren't directly developers?

(Mike Freedman at 00:41:53) We have people across lots of different areas — marketing, HR, operations, design, engineering, product. We have customer success who work with our customers. Our support people are all engineers because, I mean, we...

(Joel Beasley at 00:42:08) Oh, cool.

(Mike Freedman at 00:42:09) Have a super technical product. Our customers are engineers. And so when they're looking to speak to them, they're generally looking to speak to highly technical people, but we certainly hire in lots of different areas as well.

(Joel Beasley at 00:42:23) That's a pro. I hope you advertise that on your website.

(Mike Freedman at 00:42:26) We've really been really grateful in the quality of the people that we've been able to hire, and I think also amazed at how great people are around the world. And that's been a real boon.

(Joel Beasley at 00:42:36) Let's say one of your newer engineers, you interact with them, and I was going to say they walk up to you, but you guys are remote. You're at — you're at the company — they Slack you. I love Hyperloop. Yeah. They Slack you, and they say, hey. How can I stand out? How can I become better and be more useful? Or, like, what could I do to stand out? What would you say to them?

(Mike Freedman at 00:42:58) I think two things. First of all, is, you know, part of the important thing is during our onboarding process, we really talk to people about what we want, how people can be successful, you know, 30, 60, 90, six months. Really, as we begin to grow and put in a lot more structure around engineering and technical leadership, so have kind of dual tracks with that, but also levels by which people kind of have a clear picture about how people progress in their career, including at the IC level. Because I think a lot of engineers really still like to be deeply technical and so want to remain as individual contributors. And what that means to stand out, I mean, it's not just about how good a programmer you are, but how you can provide technical leadership, how good your communication — especially true in a remote environment — how you can get things done, various attributes which we think lead to a more successful engineer and more senior engineer.

(Mike Freedman at 00:43:52) To your question about how to stand out, I mean, I think that one of the interesting things about some people looking at their own career progression is that I think that most of the times people are so-called promoted when they've already been doing that job. So, you know, they're promoted to a certain role or level because they've already demonstrated that they're operating at that role. And so, generally, I like to think about how can you not just be passive about I'm assigned this task, can I do it? But how can you — to the extent that you're able and it's going to mean a different thing to different levels of seniority — but how can you more broadly have a positive change to your team and the company? And that doesn't only have to be purely technical if you're an engineer.

(Mike Freedman at 00:44:39) It could be communication. You know, as a developer-centric company, we write a lot of content and we talk about it, and people write great blog posts. And we get on Hacker News a lot because people are really interested in what our engineers say. We have a really technical blog. I think lots of people enjoy it. Blog.timescale.com. We also have what we call developer Q&As. So we know our users are developers. We don't put things behind paywalls, ask for your emails. We just want to put all our stuff out there.

(Mike Freedman at 00:45:08) And we do a lot of stories about what our users are doing, whether or not they're in our community or whether or not they're actually our customers. The great thing is they're really technical articles, and there's a lot of really fun things about how people are using it. And there's lots of different ways that people can contribute. And I think people are most successful when they're actually working at what they're really passionate about. And so if you want to really hone your engineering craft, do that. If you want to hone the way that you could communicate and explain to other people, do that. We've had people who move from engineering to technical product manager roles because they're really passionate about how to think about how you build better product for customers. You know? So I think that it's really leaning into your strengths while also kind of improving areas that you think you need to be kind of a balanced person. I often like to think about kind of T-shaped people. You know? People who are really strong in my area, and they try to also get balance in other areas. And that's hopefully — I work on that myself, and I think that's actually a good thing for more junior people to work on as well.

(Joel Beasley at 00:46:14) I like that. I haven't heard that before. T-shaped people.

(Mike Freedman at 00:46:17) Yeah. Usually, in certain areas, you know, you have breadth strengths and there's depth strengths. And I think it's good to try to be great at one thing, but then also have some skills in a variety of other things as well.

(Joel Beasley at 00:46:30) Exactly. Because you could be the best in the world at your craft, but if you don't have communication skills, are you the best in the world at your craft? You know? So, man, this is great, though, Mike. I really appreciate you coming on and hanging out. I mean, we made a podcast. How do you feel?

(Mike Freedman at 00:46:46) I feel great. Thanks a lot for making it available.

(Joel Beasley at 00:46:49) Thank you so much for listening. And if you found this episode useful, please share it with a friend or colleague who you think would get value from it. And if you have topics that you'd like to hear discussed on the podcast, either add me on LinkedIn or send me an email, [email protected]. Every time I get an email or LinkedIn message, it absolutely makes my day and inspires me to keep going.