Starting Your Strategic Data Foundation Now with Greg Vesper, CTO of Smarsh

(Intro Narrator at 00:00:03) Hello, my friends. Today, Joel is talking to Greg, CTO of Smarsh, and they discuss how most companies are headed towards having to deal with exabytes of data and how to set up a strategic data foundation now to future-proof your organization. All of this right here, right now, on the Modern CTO Podcast.

(Joel Beasley at 00:00:25) Here we go. This is the Modern CTO Podcast. I'm excited to talk about Smarsh and what Smarsh does. We'll get into that. But I was curious to learn a little bit more about you specifically. I thought it was really interesting that you had some history with NASA and was hoping you could share with me.

(Greg at 00:00:48) Yeah, sure. My checkered past. So I started out as kind of a young buck internet technologist working at NASA Goddard Space Flight Center on the East Coast, and this was right on the cusp of the science and technology internet becoming the commercial internet and kind of a front row seat to that happening. I had an opportunity, kind of for my first command, to bootstrap a team to create a paperless procurement system using the public internet as the actual network for purchase orders and invoices, all encrypted using public key cryptography. This was a big deal. Right? Oh my goodness, we're going to use public key cryptography and do what turned out to be about a billion dollars worth of transactions.

(Greg at 00:01:36) So we went live with this in 1993. Now you've got to understand, in 1993, there really was no commercial internet. Right? So Netscape went public in '95, which most people mark as the beginning of the dot-com boom. So we're going live with commercial transactions, B2B transactions over the public internet with some pretty cool tech, 1993. Nobody really—I think even, you know, I'm sort of amazed that NASA headquarters even let us get away with it. Right? Because the internet was not considered a commercial-grade network at the time. So that kind of lit my fire in terms of what you can do with a small team of technologists. Out of a concept, you can create software that kind of changes the way people do business, and that was just fascinating to me. And being in the NASA environment, we had access to just pretty much any technology that was out there. Right? So that created a unique combination of circumstances that allowed us to do something that was kind of the first thing of its kind.

(Joel Beasley at 00:02:39) That's pretty cool. And then back then, when all that was coming about, were people like Sir Tim Berners-Lee, creator of the web at CERN or whatnot, were they popular? Were they going to parties and like red carpets? Were they popular people or no?

(Greg at 00:02:53) No. So it's really interesting. I have in one of my presentations, I go over some of the history of this, and I have Tim Berners-Lee's first website, and it was just a little announcement. Right? On, hey, I got this new thing that I'm launching. And you don't know at the time that the world's going to change when something like that happens. It starts very inconspicuously, but then it catches on like wildfire. And I don't think it was really until when the dot-com boom started to get a lot of media attention that the pioneers of dot-com actually started to get a lot of notoriety, including one or two of my colleagues at NASA who got slurped into the private industry in a hurry. So it was a really super interesting time because you saw it coming. You saw that wave coming. You knew it was going to change the world. You had a lot of companies getting started, a lot of which petered out. But to see something coming and try to explain to the rest of the world what was going to happen, and people had absolutely no idea. I remember having a can of Coke and talking to my father and saying, everything you have is going to have one of these little web addresses on them one day that are going to point you to this thing called the World Wide Web. And he's like, what in the world? He thought I'd lost my mind. Right? But it was cool to see it coming. And then, you know, when it became kind of commercialized, it actually became less hip. It became less cool because we were the cool kids who sort of brought the internet to the business side of the house. And then when the business side of the house rushed in, we're like, we've got to go find something else different to do.

(Joel Beasley at 00:04:28) Oh, yeah. And so then what would that be today? You were able to see it then. You're active in your career today. What is the internet thing today as far as the new thing coming that in twenty years it'll be everywhere?

(Greg at 00:04:42) Yeah. So there's a couple things. I'll mention two that are sort of near and dear to my heart. One is robotics. Okay? I don't think we understand the degree to which the economics for robotics have already crossed a threshold where the components and the ability to create robotics is super compelling right now. And you see it in some specialized industries, like warehousing and the completely robotic, peopleless warehouse and some of the things that you'll see Amazon and some other crews doing. And everybody's certainly familiar with drones. But I think we're going to be shocked in five years' time, certainly ten years' time, the degree to which we've got robots running around all over the place because the economics of it are just there. I don't know that we're going to be thrilled with the result, I'll be honest with you, sociologically, but the economics of it are undeniable. So I think robotics is one that we're right on the cusp of. The other one, which used to have more of that early internet feel, it's now sort of been outed, but cryptocurrency, which is so very like the frothy dot-com nuttiness times of the late nineties. And there'll be a lot of churn that comes out of that, but you can't stop digital money. You just can't stop it. I think actually if governments of the world had understood what crypto actually represented, you might have been able to try and suppress it early on, but nobody really understood what it was. Kind of like the early internet, nobody really got it. So nobody really worked to try and suppress it or have the incumbents who control the existing financial systems. There's maybe been some attempts to do that, but it's too late now. The genie's out of the bottle. So I think crypto is really like that. It's messy right now. But out of that mess will come digital money. And once digital money is out there, you're not going to be able to stop it.

(Joel Beasley at 00:06:39) Do you hold any cryptocurrency?

(Greg at 00:06:41) I do. Yeah. I've been toying with crypto for years. You know, would that I had toyed with it earlier when it first came out and some of my crypto friends and other pioneers in technology started talking to me. I was like, okay, I listen. Fine. Interesting. I get it. I get how it works. But that was an interesting moment for me because I didn't really get it at first. I mean, I got it. I got what it was, but I didn't really get the impact of it. And then by the time I did get the impact of it, I'm like, am I getting old that I didn't get that right away? How did I not get that right away?

(Joel Beasley at 00:07:13) Oh, the amount of opportunities I turned down to be a part of crypto stuff way back in the day. But I've gotten to talk to so many of these crypto creators because I did this whole series on crypto stuff. I felt like a monkey talking to a scientist.

(Greg at 00:07:29) Yeah. There's some really bright folks in it. The basic concepts are simple enough, but it's like anything to actually take those basics and turn it into a network. Really, really fascinating stuff. So that has reenergized me a little bit to say, okay, we've got another wave of tech coming. So there's always—look, there's always the next wave of reinvention, the next wave of disruption. That's the cool thing about being in technology.

(Joel Beasley at 00:07:51) Oh, yeah. How does Smarsh fit into all of that? Where are they at in this cycle?

(Greg at 00:07:56) Yeah. So Smarsh has become a leader in our domain. And I think we ended up here because people in the very specific vertical that we're in—which is really compliance and compliance data platforms for heavily regulated industries like financial services, public sector—that vertical tends to wait until it's safe. Right? Until they move to a technology, they tend to kind of move together, and they tend to wait till it's safe. And the legacy tech in this space was all kind of built on private data centers and on-prem hosted solutions, which is an okay way to start out on anything. It's not a bad place to start. It used to be cool to deal with your own hardware, your own data centers, your racks. It was super fun. Right? And then you get to a certain point of scale where this is not fun anymore because it becomes a liability to try and keep up with this explosion of the global data sphere. So in particular, we capture every kind of human communication, whether that's email, Slack, Teams, Zoom, Facebook, LinkedIn, Twitter, text, SMS, MMS, now getting into voice, video's on the horizon. So if you think about the explosion of human communication modalities, WhatsApp, WeChat, there's no end to it in terms of the way people connect. So, A, you have an explosion of digital communications. B, the richness of those communications is going up. It's not just textual now. Now you've got pixel data, voice data. You've got big data kind of growing at least geometrically, some would say exponentially. So you got this explosion of content. And in the regulated industry, your retention periods are five, ten, fifteen, twenty years. So all of a sudden, let's say you started out building your own little local data store and you've got your racks of equipment and your RAID arrays, and let's say like, oh, I got this cool kind of thing going. All of a sudden, you're into a multi-petabyte domain, and exabyte is on the horizon, and you have to have a geo-redundant data store, and you realize that you don't really have the ability to build that out economically. And that's one of the things that's changed now, is because of the scale in cloud, the economics of geo-distributed compute and geo-distributed data stores has shifted in favor of the trillion-dollar infrastructures. They have the really good unit economics. And if you're trying to build that out yourself, you don't have the purchasing power. You don't have the footprint. You don't have the ability to build that out at scale. You can build out anything if you're at a modest level of scale. But as soon as you get into this multi-petabyte space, you find yourself kind of stuck.

(Joel Beasley at 00:10:44) Interesting. And so comms is really broad. Where do your customers sort of group? Give me a couple examples of companies and why they're storing their comms and what type of comms they're storing.

(Greg at 00:10:58) Well, it started out—this whole industry started out with email. Right? It was email archiving, and there are FINRA and SEC regulations that say, hey, if you're in a regulated space, you have regulated users, you're dealing with financial transactions, all of this has to be retained for e-discovery, litigation, audits, and just basic regulatory purposes to make sure no one's breaking the law. Right? So simple enough, email, and then things start to evolve. And you get other communication channels that come on top of that, instant messaging, text, MMS. Then what you have is the clients of these financial services institutions have their own mobile phones and their own way of communicating. And there are sanctioned channels and unsanctioned channels, and a client comes in on an unsanctioned channel, and the bank says, I'm not going to not talk to my customer. Right? And tell them, well, we can't do business with you because you use WhatsApp, WeChat. Those are unsanctioned channels for us to chat. So the regulatory envelope keeps getting bigger and bigger, and then the companies involved have to adapt. And that's what really created the opportunity for Smarsh. Because Smarsh really handles three pieces of the equation. One, the full-fidelity content capture, meaning you need to get all the data, the contextual metadata. You need to understand kind of everything that you can about the communication that happened. So we do the full-fidelity capture across all of these different communication platforms, including emerging ones. Then we do the retention over arbitrary periods of time for retaining this data, and then we do the analytics on top of it. Because if you think about the regulatory problem, look at the conversation you and I are having. We'll talk about all sorts of different things. How much of it is of regulatory value? Very, very little. So you have a needle in a haystack problem. Let's say that we were talking about a trade and you were a customer and that was a bank. We might talk about a bunch of things and there might be a very small snippet that actually had regulatory merit. So you now have this sort of needle in a haystack problem, and the haystacks are getting bigger. So you've got to have some really intelligent tech for being able to find the needles in the haystack. So that's really the three domains of our business. Right? The full-fidelity capture, the retention, and then the analytics to find the needles in the haystack.

(Joel Beasley at 00:13:09) So email, text, those seem pretty obvious because you're first-party in control of them. But for WhatsApp specifically, do they have an API for that? How do you actually get that data?

(Greg at 00:13:21) Yeah. That's one of the more interesting ones. Okay. Because with WhatsApp and WeChat, there are versions where there are publicly available APIs, but that's only a subset of what they offer. Because if you get certain communication channels like WeChat, they're designed to have secure, encrypted communications that you can't monitor. You can't intercept. You can't do a kind of an actual violation of that session. So some of them are harder than others. And when you have to do that sort of man-in-the-middle proxy, it's challenging technologically, and you're always wondering about the terms of service on the one hand, and you tend to end up with solutions that are a little bit brittle. Because they can change the protocol, and then all of a sudden your proxy breaks.

(Joel Beasley at 00:14:08) Oh, yeah. Well, my first thought was I was in the financial services industry before all the amazing tools today existed. So we would have to do all of these crazy workarounds to screen scrape, and we built all these systems and all of these things to get the data out of their accounts. We'd log in, download PDFs, extract it. And so when I was thinking about WhatsApp, I was like, how would you do that? There's so many ways running through my mind of things that you could do.

(Greg at 00:14:32) There's a lot of different techniques that people use to do it, but they're all various different flavors of brittle.

(Joel Beasley at 00:14:39) Right. Exactly. And so then you have to have teams and then maintain those and build in—

(Greg at 00:14:43) Then it's a sustaining engineering problem. That's exactly right.

(Joel Beasley at 00:14:46) Yes. Exactly. At least it's tied to compliance, which is something people have to do. So the market is there. Because if it was a much smaller market, it'd be a lot harder.

(Joel Beasley at 00:14:56) Yep. So how did you even get involved with all of this in Smarsh? How did you meet the team there?

(Greg at 00:15:02) It was purely by accident. I had exited my last technology engagement after a private equity transaction. I was taking some time off and motorcycling up the West Coast. And a buddy of mine who was a CEO in the K1 portfolio said, hey, you should meet with these guys. So I met Sujit, one of the K1 principals. He runs operations for K1. And Sujit lives in Portola Valley, and I zipped through on my motorcycle, and we hung out for coffee. And we're just chatting it up, and he started to tell me about this merger of two companies in the compliance domain, one an SMB mid-market company, and the other enterprise. Right? In the same vertical, the same solution space, but the tip of the pyramid versus kind of the fat middle and the bottom of SMB and mid-market.

(Greg at 00:15:57) And they were about to put these two companies together. One was Legacy Smarsh. The other was a company called... And they were looking for some help on kind of product strategy and technology. And it was all very interesting.

(Greg at 00:16:10) I'm, you know, in my motorcycle garb and have a week's worth of growth on my face. I might have even had hair back then. So I said, yeah, you know, it's interesting. And so by the end of the conversation, Sujit says, so what's your availability?

(Greg at 00:16:24) I was like, well, see that motorcycle over there in the parking lot? That's mine. And what I'm gonna do is get back on my motorcycle and finish my trip. In fact, what I ended up doing was parking my motorcycle in the Bay Area, flying back home, and getting involved because this acquisition was right on the cusp of happening. So I got pulled in a lot sooner than I thought. And that's when I met the then CEO of Smarsh, Brian Kramer, and we hit it off, and on we went.

(Greg at 00:16:51) It's all about timing. These things are sort of serendipitous.

(Joel Beasley at 00:16:55) Yes. And often, you know, we talk on the show or people talk to me outside of the show about, you know, relationships and the importance of them long term. This individual that pulled you in, can you tell me a little bit more about how did they come to the thought that you would be a good fit? How did they know and trust that you would be right for this?

(Greg at 00:17:17) I think it usually is based on relationship to a degree. Right? And who are the folks that you've been in a foxhole with? So Sujit, the K1 principal I met with, had a relationship with this CEO, this portfolio CEO. And Sujit, after talking to me, calls this portfolio CEO and says, hey, will you vouch for this guy? Does he have what it takes to get the job done?

(Greg at 00:17:36) And this portfolio CEO is a longtime colleague of mine, and we did some great work together in the B2B space. And so, you know, when you get that kind of resonance where you've got people whose work you respect and whose character you respect and that network effect. I think that's probably the way most of these placements work.

(Greg at 00:18:02) And it's not that there isn't a place for executive recruiting, and certainly that happens. And certainly I've had positive experiences there as well. But it's really the things you weren't planning on. I had no intention of starting a new gig when I got involved with Smarsh, not at that moment, anyway. I was just kind of airing my brain out.

(Greg at 00:18:20) But, you know, things happen. Electrons fire, paths cross, and then all of a sudden you're back in pocket. And that was coming on five years ago now.

(Joel Beasley at 00:18:30) Wow. What's been the big changes in the past five years since you joined?

(Greg at 00:18:34) Yes. Smarsh has changed a lot actually because we have grown both organically and inorganically, and we saw some white space. And this kind of goes back to our earlier conversation on what's kind of unique about Smarsh and what is sort of internet-like in its innovation in this domain. And in this case, the obvious innovation that the rest of the competition and the industry as a whole had not embraced will not seem terribly innovative, but it's public cloud infrastructure.

(Greg at 00:19:06) Now in the compliance domain, public cloud, until the past couple years, had a scary and inappropriate ethos for the compliance folks to take because it's public. The word public does not help the public cloud providers when they're going into the compliance domain because it sounds like it's public. Right? You know, anybody can get access to my data.

(Greg at 00:19:28) So when you get into those kind of inherently multitenant infrastructures, right, and heavily virtualized infrastructures, there was, I would say, some caution on the part of financial institutions in terms of getting involved in that kind of infrastructure. But Smarsh saw some white space there and said, look, nobody has actually built a cloud-based compliance platform. Right? Nobody had actually done that. So we went down that path and actually built out, you know, kind of a multi-cloud path that would allow us to, you know, multiple public clouds and scale horizontally with very aggressive unit economics.

(Greg at 00:20:07) So it was a little bit of an odd experience for me because when you get involved with a new space, I'm always kind of trying to look to my right and my left and over my shoulder and say, okay, who are the fast movers? Who are the competitors in this space? As we moved into cloud, there was nothing on my left, nothing on my right, nothing behind me. And at that point, you're like, okay, there's probably a reason for that. Why aren't other folks kind of rushing in?

(Greg at 00:20:31) And I think people were not prepared for the shift in the global data sphere explosion. And where private cloud and private data centers were still, I would say, the de facto safe choice for financial services for compliance platforms. And I think it snuck up on everybody that you're gonna have this explosion of content. So by the time we had actually deployed our public cloud platform, I would say we might have been six months early to market, maybe nine months early to market.

(Greg at 00:21:05) At first, you're like, oh, jeez, should we innovate too soon? Which has been my weakness as an innovator is to be in front ahead of the market. And people might think that's a good thing, but it's not a good thing. You don't wanna be ahead of the market.

(Greg at 00:21:18) All you end up doing is you deploy a bunch of capital. You educate the market. Somebody else comes along and says, that's a great idea. They hit it with the exact right timing. So I was a little concerned at first that maybe, you know, I jumped to the next curve here too quickly.

(Greg at 00:21:31) I think we were maybe six to nine months early, but then all of a sudden, it became really clear that folks who were doing private data center or doing on-prem archives, they're all hitting the wall in various ways. And that it just became a scale issue. It's like there's no way to scale it. So our innovation in the space was really to say, we're gonna take compliance platforms to the public cloud. We're gonna bet on that.

(Greg at 00:21:53) And I would say our timing was just about right, maybe a tad early.

(Joel Beasley at 00:21:57) And so today, are most of your customers coming from people who had rolled their own product and they're hitting a scale now in which they need to, you know, offload that to someone else? Or is it companies that are just growing to the point well, they... I guess they've always had to do it. If they're in a compliant-related industry, they have to do it basically from day one. Right?

(Greg at 00:22:17) That's correct. Everyone has a data problem that is accretive. So it's a question of where does it live. The folks who are in the most pain are the folks who built this out on-prem. Because when you start to get your own kind of infrastructure footprint, and remember, this has to be geo-redundant.

(Greg at 00:22:35) But it's one thing to build out a footprint. It's another to say, okay, I'm gonna have, you know, an active-passive topology where I pay twice for something I only use once. And then if you try to actually get, like, triple redundant, you're done. Your economics are toast. You can't build it out yourself in a cost-effective way.

(Greg at 00:22:48) So the people on-prem are the ones who are hurting the most. The folks who are using kind of vendors with private data centers and, you know, their own sort of brick-and-mortar data centers, they're getting a little bit further. But I think what happens there is they start to realize that, well, wait a minute. I'm captive.

(Greg at 00:23:07) I'm now captive with a multi-petabyte problem in someone's private data center. And if I ever have to move it, it's going to take years. And when you start your planning horizon becomes years and realizing that, you know, again, the data sphere problem is only going up faster. All of a sudden, you realize, wait a minute. I gotta park my data once and forever in a place where it can grow.

(Greg at 00:23:31) I've got good unit economics. I've got good information security, and I don't have to keep thinking about migrating it. I have to land this plane at one point. And that's where, really, the trillion-dollar infrastructures are the only safe place to actually end up landing that data. So I think everyone's cycling around to that now.

(Greg at 00:23:48) It varies, but the folks on-prem are the ones who are feeling the pain the most.

(Joel Beasley at 00:23:52) And so if there are any compliance-related folks on-prem listening to the show right now, what would the next step for them be?

(Greg at 00:23:59) Well, so it's really just all about your data trajectory. Right? You have to do the math and say, how much data am I ingesting? How much data am I retaining? Now some of the folks that we work with, the retention challenge is aggravated by the challenge with actually purging and deleting old data.

(Greg at 00:24:19) Some of these older systems, because these archival systems are designed for retention and because they're designed to ingest quickly, they're not designed to delete quickly. You actually have a fair few number of folks who are retaining data simply because they don't have enough horsepower or enough certainty to be able to delete. So what you will get is this problem is kind of growing accretively, and there's no real way out of it.

(Greg at 00:24:59) As a matter of fact, we have folks who move data to our platform just so they can get enough clarity in order to do their deletions. Right, and their purging, and then kind of move the platform forward. So that's perhaps an extreme case, but not uncommon. It's actually fairly common that you have these archival platforms where deletion or what's called disposition is challenging.

(Greg at 00:24:59) So for those folks, I would say, look, the longer you wait, the worse it gets. That's not meant to be scary. It's just meant to be kind of brutally honest. If you have a data gravity problem in your infrastructure or a data gravity problem in someone else's infrastructure, the longer you wait, the worse it gets.

(Greg at 00:25:25) So there are some simple techniques. Cut over to a strategic place. You don't necessarily have to migrate all your content. If you can just cut your day-forward data over to something strategic, then you can age out and delete out, and you don't necessarily have to do a big multi-petabyte migration. You can, and we have folks who do both.

(Greg at 00:25:45) We have some folks who say, hey, look, I'm gonna migrate everything. We have some folks who say, hey, I'm gonna be split-brain. Right? Yesterday's data is over here. Tomorrow's data is over here. I'm gonna age out. And maybe once it's small enough after aging out, I'll take the remnant, right, and I'll pop it over once the lines kind of cross in terms of the economics of moving that data over.

(Greg at 00:25:56) But I think the most important thing is get a strategic data foundation now. Don't wait. The bigger the problem gets, the harder to solve, the more expensive it is to solve. And the economics of it are just, they're not overwhelming to do it. Right?

(Greg at 00:26:15) Just get started with your day-forward data. And then when you... you can decide how you wanna handle your multiple petabytes, but everyone, at least the larger institutions, need to understand they're headed toward an exabyte problem domain. Right? So if you're heading toward exabyte problem domain, get yourself a nice foundation to build on, and then, you know, you can work your way there over time.

(Joel Beasley at 00:26:38) Okay. Cool. As I go about my life, if I wanna casually bring stuff up to a CTO and technology leader about what you guys do, should I just be like, hey, you should just check out Smarsh and take a look at what they're doing in the compliance space, or is there any sort of thing I could say that'd be useful to catch their attention?

(Greg at 00:26:56) Well, I think the why Smarsh is really what's unique. And what's unique is that cloud-based platform for compliance that kind of makes it easy to get started. We toy around with nonregulated verticals occasionally. Right? Folks who are saying, hey, I'm not in the regulated space, but I do wanna be able to retain my data, and I do wanna be able to find needles in haystacks for other reasons.

(Greg at 00:27:12) I would say, you know, take a peek at your data trajectory and ask yourself how important your data is to your business. You're... I'm talking about communications data now, not transactional data. This is, you know, strictly human comms. And what we're finding for some folks is that over and above regulatory use cases, there are use cases where people wanna understand the sentiment and behavior of their interactions with their customers.

(Greg at 00:27:45) So can you do sentiment analysis on... yeah, you can. You know, you can say, now that I'm looking for needles in haystacks, now that I have, you know, a combination of technologies that let me actually do kind of some proper AI analytics on this content, then it might become pretty interesting to say what other behaviors, even nonregulatory behaviors, are meaningful to your business. What are the attributes of a particularly successful sales rep?

(Greg at 00:28:11) Right? It'd be interesting to know. So it can take you in a lot of different places. We tend to focus primarily on the regulatory use cases and secondarily on the behavioral use cases. But depending on the folks you're speaking with, if it's regulatory, it's a no-brainer to take a peek at Smarsh. If it's nonregulatory, then it's a maybe depending on what you're looking for.

(Joel Beasley at 00:28:30) Pretty cool. And those things like sentiment analysis, are those things that you guys build? Are you API... open APIs to let them build? How does that work?

(Greg at 00:28:38) Yeah. So that's a really interesting question because a lot of our machine learning tech, we incubated in very kind of bespoke deploys with a tool-based approach where we provided the tools and worked with enterprises where their data scientists and their engineers would, you know, try to build the right model for finding a certain kind of behavior. We learned a lot through that.

(Greg at 00:29:26) First of all, we learned that the vast majority of machine learning deployments in the enterprise space are tech-enabled services to help solve a bespoke problem, which is not a bad way to learn, but it's also not a great business model. And if you look at a lot of the players in this space, there's a lot of bespoke tool-based, tech-enabled services out there.

(Greg at 00:29:26) So we looked at that and said, look, it's interesting. You'll learn a lot, but what we really want is to productize this machine learning capability. So one of the innovations that Smarsh has is that we've taken our learnings and embodied those in our own proprietary machine learning models and proprietary artificial intelligence analytics. And we have so much experience with this communications data and these specific regulatory use cases that we're able to package them up in a way that is actually productized and reusable across customers.

(Greg at 00:30:13) That's where you really crack the nut on value in terms of a business driver anyway because you have to be able to create leverage, and the leverage comes from reuse and repeatability. And if you've gotten deep in the data and you know the models that know how to find what you're looking for in that data, so we now package these things up into what we call scenarios or intelligent scenarios where we combine a number of techniques, not just machine learning, but lexical scanning and data classification and machine learning all combined, right, into one intelligence engine that says, I know what you're looking for.

(Greg at 00:30:55) Right? If you're looking for a gifts and entertainment violation in this kind of communication channel, this kind of modality, this is how, you know, the semantics of how people converse. Because remember the semantics of an email versus the semantics of a text versus the semantics of a voice conversation, very different. Right? You have very different words that you use.

(Greg at 00:30:55) So if you kind of start to do that semantic mapping with some of your technology and get to say, hey, for this modality, this kind of behavior you're looking for, we have an intelligence engine that can find that. And that, I think, it goes from machine learning as a hobby or a science experiment into machine learning as a kind of reproducible business model.

(Joel Beasley at 00:31:21) Yeah. That sounds like you guys have a lot of fun over there in the tech department.

(Greg at 00:31:24) It's crazy. It's crazy what we do, but, you know, it's fun. It keeps us busy.

(Joel Beasley at 00:31:29) As we wrap up, I want to talk a little bit about leadership. Obviously, the company's large, it's growing, doing really interesting, what I personally think sound like fun things. What do you tell the next generation, the leaders that are coming up, people who you're speaking to within your company? Like, how to stand out, or what you look for behaviorally or culturally so that they can get ahead in their career?

(Greg at 00:31:56) Yeah, great question. We could do a whole session on that one. So leadership really matters, and it's variously defined. But to be a leader, and particularly in this space, first of all, technology is kind of the science of change. And if you want to really be good at it, you have to change frequently. Right? So if you really want to get good in tech, you've got to really kind of ride that knife edge of the change and the change management. And if you have a static mindset... I think when I started out at NASA, I had a lot of energy and a lot of intensity, and I was a bit of a perfectionist. Right?

(Greg at 00:32:43) And perfectionism actually can work against you. And I had an excellent mentor there who saw that my perfectionism was holding me back. And he said, "Greg, make your mistakes. That's how we learn. If you want to be perfect," and this is how he got my attention because I wanted to be perfect, "If you want to be perfect, try not to make the same mistake twice." And that pierced my skull. Right? I completely got what he was saying.

(Greg at 00:33:14) So you got a very dynamic industry where there's a lot of adaptation, there's a lot of experimentation. And leading means getting really good at that experimentation and adaptation, because at the end of the day, it's empirical. What we're doing, even in software, is very data-driven, empirical, and you have to get really good at being able to suss out what has value and what does not have value. And if you're going to lead, you can't just execute. Right? Managers can just execute. If you're going to lead, you need to know where you're going. Right? And the only way you're going to know where you're going is if you're plotting the trajectory and trying to say, "Well, what's the value point?"

(Greg at 00:33:56) And I think in technology, people do confuse leadership with execution and say, "Look, I'm really good engineer. I really know how to execute. I can solve any problems." Like, okay, awesome. Leading means knowing what problem you want to solve. And how do you assess that? And that has to be done in the context of the marketplace. Right? Because we are all working for the marketplace, and you can have the greatest ideas and the greatest technology. And if no one will cut a check for that great idea or great technology, then it's either education or research, but it's not actually going to be a career builder.

(Greg at 00:34:32) A career builder is the intersection of that technology and market traction. And so I think all up-and-coming leaders in technology do need to be a little bit of a product manager, a little bit of product strategy, and understand I've got to have strong product-market fit. If I don't have strong product-market fit, I can do a lot of cool stuff. I've done a lot of cool stuff. I did a lot of cool stuff in the robotics space early on with some drone tech that was obviously going to change the world. And it did, but it wasn't my company that changed the world. But I had the tech, I had the tech for it. Right? So you've got to get that product-market fit, and you've got to get the market timing just right. That becomes the better part of leadership, because if you don't end up with a successful economic model, then the technology becomes irrelevant.

(Greg at 00:35:20) And one of the hard things, I think, for technologists to learn as they get into this is that you need enough technology, just enough. It's not all about the technology. It's all about the product-market fit, and you need just enough tech and then continuously churn and churn and churn once you've got that traction and that impedance match with the marketplace. That really is number one. Then there's a lot of other domains, pieces of leadership in terms of team building and execution and methodology and that. But I think number one is make sure you're building the right thing.

(Joel Beasley at 00:35:54) I love it. And I know we're coming up on time. I think that if you're into it, we should consider having you on for like a pure leadership podcast. When I come across people like you, I'm like, alright, we've got to have them back on.

(Greg at 00:36:05) Let's do it.

(Joel Beasley at 00:36:06) Do a pure leadership episode because you've got some really good ideas.

(Greg at 00:36:09) That's one of my favorite topics. And there's such a need for it. So I've got a lot of passion on the topic and we're all students of leadership. So happy to share the things I've learned today, and I'm sure tomorrow I'll be learning a few more.

(Joel Beasley at 00:36:22) Perfect. Well, thank you so much. We made a podcast, Greg. How do you feel?

(Greg at 00:36:25) Outstanding. Thank you so much.

(Joel Beasley at 00:36:28) 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.