Andrew: Hey there, freedom fighters. My name is Andrew Warner. I’m the founder of Mixergy where I interview entrepreneurs about how they built their businesses. Joining me is Leon Kuperman from cast cast goal is to cut your cloud bill in half. And I invited him here to talk about how they do it and how.
And I also wanna find out what happened to his eye that time after he sold his previous company. We’ll talk about that and so much more. Thanks to two phenomenal sponsors. The first, if you’re hiring developers, go to lemon.ai/mixergy. The second, when you’re ready to do, um, marketing automation, right? Go check out, send in blue.com/mixergy.
Leon. Before we get into cast, you sold your previous company, Zen Zen edge cybersecurity company to Oracle. I couldn’t find out how much, how much did you sell it?
Leon: Wasn’t publicly disclosed,
Andrew: it wasn’t
Leon: a good.
Andrew: was it? I feel like if Oracle, a publicly traded company is not revealing the number, it’s not significant enough. And I, I, all I’m seeing is you didn’t raise much money with it, and then you sold it to Oracle, but I don’t see for how much was it a good exit for you? Life changing money? I
Leon: Um, it wasn’t, uh, life changing is a relative scale at the time. Yes. At the time it was life changing. but it wasn’t the kind of life. It, it wasn’t like, all right, that’s it. Uh, I gotta retire, like, fortunately we’ve had a couple of other good business successes since then, but, uh, but, but XGE was an interesting experience.
Cause it was the first time I had a business that was acquired by a large, a large company. And I’ never worked for a large company up until that point in my career. Really the last one was IBM and that was when I was a student and an. So this was a really interesting ex uh, uh, experience for me to get integrated with big red.
Andrew: How long were you with them? A couple of years. I think
Leon: Yeah. But, uh, almost three years. And what was really interesting there is I thought I would be there for kind of a, a rest invest situation. And, and then Oracle has this really especially OCI, which is Oracle cloud infrastructure, which is kind of a, a Maverick division within Oracle. They don’t behave.
Like the rest of the company, they really have interesting ways of vetting out talent and then getting those people to step up. So they found me and they said, no, no, you can take a much bigger role. And then I ended up getting an apartment in Seattle and then ran. So Xanax was just one security product.
And they had a whole suite of them. So I ended up, uh, running the entire, it was 200 and some odd person organization with multiple service teams. And what was really cool about that experience is I got a chance to pitch my roadmap to Larry Ellison. Uh, and he was super engaged. It wasn’t like. You know, a 70 year old dude that was no, no, this guy was on it.
And I was really impressed by, he was almost my biggest cheerleader in that roadmap meeting. So it was a lot of fun that contrary to my earlier bias, it was a great experience.
Andrew: because what, what did you learn from it? Or get out of it that made it worthwhile. I’ve had friends who rested, invested. It’s a really nice experience. What did you have? That
Leon: Yeah. So we could talk about that. I don’t like to rest. I, I don’t enjoy, I enjoy what I do. So for me, I, I get to work and then do what I love at the same time. So twidling, my thumbs is kind of painful, but. Here’s what I learned to answer your question explicitly, Andrew, I thought of myself as a pretty strong writer, uh, and communicator.
And what was interesting is that Oracle, the team was infused because of the Seattle location. We had a lot of AWS specific infusion of talent. So a lot of the senior leadership, a lot of the engineering talent. Came from AWS and AWS and Amazon in general has a very different approach to meetings, approach to writing.
They use this thing called a P R F a Q and at first that was very reluctant to learn anything kind of, Hey, I’ve been writing for 20 years. Don’t teach me to write, but when I got into it, I realized this was a very crystal clear way of communicating intent and outcome. What’s considered a successful outcome for a project.
And I’ve adopted it since then. And I use it to this day and I, you know, teach my team how to write in that style. And it’s been extremely effective
Andrew: you mean? Can you gimme an overview of what it means to have the outcome be clearly articulated in this process?
Leon: yeah. So Amazon uses this thing called the P R FFA Q. Um, it’s basically a product document, but they start with, before they write a line of code, they start with a press release. You know, Amazon today announces I’m gonna make this up relational database service that does and outlines the benefits and how the customer interaction and, and the customer experience.
And then the set of FAQ questions that fill out the story. So it’s as if you’ve built it and you are articulating the final outcome to the market. And I thought that was a really insightful way of describing the product.
Andrew: Got it. It’s PR for press release, FAQ for frequently asked questions. You have the combination before you start. And that way you’re keeping the end in mind. And I imagine you’ve heard about them doing this, but it wasn’t until you started to push yourself or get pushed into creating that document that you realized, wait, this does make sense.
And now it’s part of what you’re doing. A cast.
Leon: Yeah. And I had to, in order for me to build a roadmap, an engineering roadmap and a product roadmap at Oracle, I had to get through three or four levels of review, um, with senior leadership before I could actually pitch it, it went all it all. It goes all the way back to a PowerPoint form for, for the most senior management in the company.
But that whole process was pretty enlighten.
Andrew: All right. Um, before we get into what happened at CAS and how you ended up creating this company. You told me that once you sold, you said I’m signing up for what?
Leon: So I’ve been doing this, uh, martial art called Brazilian jujitsu for a long time. I actually started in martial arts because I was a skinny kid that kept getting picked on in public high school. Like, you know, parents couldn’t afford to send me to, to private.
Andrew: what, what would your parents do to you? I mean, what would the kids in school do to you?
Leon: Probably the worst. And, you know, nowadays this is unimaginable, but probably my worst incident was it’s not just like a schoolyard fight or I had the class bully, put me in a headlock for all of his goal was to see if he could hold me in a headlock for all of lunchtime.
Andrew: Wow,
Leon: an hour and whatever
Andrew: humiliating too, that everyone else gets to see that it’s not just the physical part of it, that everyone gets to see you not be able to leave.
Leon: yeah. And forget about getting like stuffed into lockers and all that stuff. It was pretty BR like, you know, like think about back in the early, in the mid eighties or early nineties, like, you know, it was a brutal bullying environment.
Andrew: was encouraged almost. It was supposed to make a man out of you. It was supposed. Teach everyone else that they have to stand up for themselves. There would be these lessons from fathers and from movies about how it’s your fault, because you don’t know how to stand up for yourself. It’s the dad’s fault for not teaching you how to punch.
And so you went through that. Did you internalize it that way? Did you say there’s something wrong with me? I’m a loser here. You did.
Leon: Absolutely. Absolutely. And, and so in grade 10 or some something like, I think in grade 10, I saw, I used to be a big, I think for that reason, I used to be a big profess professional wrestling fan. Like I just loved the simplicity of good versus evil. But when I got to high school, I saw a real wrestling match, like a freestyle wrestling match.
And I’m like, Hey, if I did this, maybe I wouldn’t be like, I couldn’t win, but at least I would understand how to physically protect myself. So I signed up for wrestling and then that’s what kind of many years later brought me to Brazilian jiujitsu in Los Angeles. so, so I, I was, uh, at the time of that Oracle acquisition, I think I was a purple belt or maybe a blue belt.
Yeah, no, I was a purple belt. Purple. There’s only five levels in jujitsu. It’s white, blue purple, brown, black. And it it’s very, it depends on how much time you put into it, how much you compete. What your attitude is, it’s, it’s really a lot of, there’s no formal grading system. It’s up to your professor in, in Portuguese.
They call them professors. And so I didn’t dedicate a lot of time. Maybe I’d go once a week, once every two weeks. And my progress was really slow, but at the end of this kind of acquisition process, and we worked really hard to get the deal done. Um, I said, you know what? I can now. Kind of sit back for a minute and I wanna go and compete and see how good I can really get.
And so I dedicated myself to training every single day, sometimes, twice a day. Um, and, and that was kind of a moment of acceleration for me because I could really apply myself to a skill that I wanted to get good at.
Andrew: And so you, you said I’m signing myself up by the way. I, I need to do that now. I told you I moved to Austin. It’s so hot here. I’m not doing as much running as I used to. I think I need to sign up for a marathon to really give myself a goal that I have to work towards. Otherwise I’m just gonna, I don’t know, not push myself nearly enough.
And so you said I’m gonna go in into a fight. Is it called a fight?
Leon: It’s a match. Yeah. There’s like a. Couple months down the road. Uh, I think it was, uh, I think it was, it was called the PanAms at, at that, for that particular tournament
Andrew: And so you train to get there. How often,
Leon: every day trained every single day.
Andrew: every single day by yourself, or you go in and get a coach.
Leon: Yeah. So my coach, so I, I was already part of a club. So my, my problem was frequency. I just didn’t do it often enough to get really good at it. So, but they, they run classes every day. So I started, uh, I started doing classes every day and I got myself, a private coach. So the combination of the two private instruction.
Which I could afford to do now, you know, it was a, it was a possibility. And the fact that, uh, I was super dedicating myself and I, Andrew, I think I told you like one of the first days I think it was a Monday, uh, I was trading at lunchtime and someone hit me in the, uh, upper, upper eye and it split open pretty nicely.
So I take like a selfie. And my caption was another day at the office, cuz I really felt for me that was freedom. That was my personal freedom. Right. The ability to train when I wanted to.
Andrew: Yeah, I get that. All right. And so then you said, you know what, I’m gonna start another company based on a problem that I experienced before. And the problem as you told me was at one point, your AWS fees were just a couple of thousand dollars a month, right?
Leon: when we first started Zenge they were really low because we just had basic infrastructure.
Andrew: Okay. And then, and we should say AWS, Amazon web services. These are the computers that Amazon keeps in the cloud for companies like yours to use, to run their businesses. You’re paying them a couple of thousand bucks a month, which is considerable, but not, not that significant. Then it got to how much the.
Leon: So by the time we had, um, that, uh, consummated a deal with Oracle, it was closer to $300,000 a month with AWS.
Andrew: So that doesn’t seem like that much though. I mean, don’t get me wrong. It’s a lot of money, but you’re looking at a company that raised millions of dollars that sold to Oracle that had, um, actually, I don’t even know if it’s a lot of money. It just doesn’t seem like that much in, in comparison to what I imagine you sold for.
Why was that an issue for you?
Leon: well, it was, you know, we’re a very cost conscious company, 3.6 million a year. That’s a big portion of our overall burn.
Andrew: Mmm.
Leon: you know, we were, and, and, and so here’s why it really matters. Andrew, when you’re a SaaS company software as a service, um, you’re valued based on your gross profit margin. So that is what you sell the software for minus your cost.
And if cloud costs are eating up into that gross margin, your valuation is going to drastically suffer. So anything worse than an 80. Percent gross margin you’re valued differently. And so we were very concerned, especially since we were hitting our hockey stick growth curve. Where, what does that happen when we 10 X again, and at that level, that becomes a significant problem.
And I can only, I can point you to a public example of that problem magnified significantly a couple of weeks ago. Um, Snapchat reported their public earnings and you know, it wasn’t great, but here’s the problem they’re hosted entirely in the public cloud between Google and AWS and to solve their gross margin problem.
They prenegotiated discounts with both Google and Amazon, and they said, look, we’ll pay you guys a bunch of money up front, but give us a significant discount. And so they did that horse trade. They paid AWS a billion dollars and they paid Google $2 billion. And they’re still paying off that schedule.
It’s a five year program. The problem is nobody can predict what’s gonna happen to your business in five years. Maybe the Snapchat team didn’t think about a recession at that time, but how can you, like, you know, how do you plan for that eventuality? If you’re locked into these costs? For a five year period.
So that’s an extreme example of what the microcosm that we were going through.
Andrew: And what you told me before was, you said, look, I could have cut some of it, but how do you know what to cut? What if we cut some of these expenses, some of these resources, and then it turns out that that’s exactly what we need in order to stay up, to provide the service that our customers are paying us for.
Am I right?
Leon: Yeah. And so it’s not so much that I don’t know what to cut at the moment. It’s what happens when traffic surges unexpectedly in 10 minutes from now or an hour from now or over the weekend? When I don’t have my DevOps engineers, they’re just on call. They’re not, they’re not sitting in front of the helm.
So the problem was predictability on of traffic and we just didn’t have the tools to foresee every single blip on the radar.
Andrew: All right. Let me do my first sponsor. Uh, I’m sorry, I’m gonna do a first sponsor. I’m gonna do it in three sentences or less. Let’s see if I could do it. If you’re hiring a developer, go to lemon.io/mixergy. Lemon matches you with phenomenal developers at a reasonable cost. that’s it. lemon.io/mixer G I did it.
So you are, you were starting to say that that’s the problem about how do you know going forward, especially what you cut, especially when you can’t predict what traffic’s gonna be like.
Leon: Yeah, that, and, and that was in, and that was in our case. We. Ended up going through a bunch of motions, like tried to reason through, well, why don’t we do similar discounts? Kind of what I mentioned, the Snapchat example where you commit a whole bunch of money up front. And we actually committed like close to a million dollars.
I think it was $750,000, which was a lot of money for us, but it’s something that we felt that we had to do. And then, so you go to, it’s almost like grief counseling. You go through all these faces of. When you’re Andrew, have you ever seen an Amazon web services invoice? Have you, have you ever put your hands on these.
Andrew: Yes, it makes me feel like such an idiot that I don’t understand it. And then I, this was actually in the beginning when I went in, it made me feel like an idiot for not understanding it. And then I started discovering that there were companies that would help explain the bill and start to figure it out.
And then you have to hunt and see, at least in my case, we don’t need it that much. In my case, I had to go and figure out what of this do we actually need? What do we not? And it was stuff would just get stored because it’s just in case, let’s make sure we have it, but then we didn’t really need it. And it became a liability.
That’s that was my experience with it. What did you see?
Leon: Yeah, it’s the same thing. All customers feel silly. You, you have to be a rocket scientist. The way those bills are structured is every single fractional component that you use is broken down into its online. It. It’s impossible to understand that as a human being. And it’s impossible just to understand it.
So there are companies, there’s a whole ecosystem of companies that have been built and sold. Now that help you understand the bill. But my, my, my hypothesis is with my team, is that understanding it doesn’t do much for you. That’s just understanding it. How do you fix it? Like how do you get to the root of the problem and have a systematic approach to fixing how much money you spend?
And in a way it’s such an existential problem, not only for customers, but for the cloud industry in general. And I, and I wanted to dig into that with you, if you, if, if we have a few,
Andrew: yeah. Do
Leon: So, you know, right now cloud adoption is growing at 40% year over year, Google just grew 40%, roughly the same for the others.
So nobody is short on business, but what happens when we get to kind of the laggard end of the adoption life cycle of the adoption curve, and we start getting even the healthcare industries in the cloud, you are going to have in order to get there, growth is gonna slow significantly. And the reason it’s gonna slow is because customers are gonna say, wait a minute, this is way too expensive.
Maybe my old data centers. Are better economically for me than the cloud. It’s this whole movement called repatriation. And that’s what AWS, Amazon web services, Google and Microsoft Azure need to be worried about. And I think that’s what they are worried about. Existentially in
Andrew: the way, this is what you told me, you were thinking of doing you. What this is too expensive. What if we just buy our own hardware, put it in, uh, racket ourselves and start to manage it ourselves instead of paying for them. Am I right?
Leon: Yeah, we tried. I mean, it was a terrible idea, but we tried it and like, that’s the extent that you go through, like, we had 30 points of presence, what are called pops. They were one rack unit. So like what a, however many servers you can fit on a single rack. In a data center, which included all of the networking gear and the transit gear, like basically it was a self-contained bundle and we, we built 30 of those out and spent quite a bit of money doing it.
Um, and that was even, I mean, it was cheaper to run a baseline, but we found ourself going back to the cloud for elasticity. Like once we maxed out rack. It’s really slow to buy another one. it takes months to buy more equipment. So you don’t have any elasticity in that environ.
Andrew: And still you’re saying, look, at some point, some companies are going to go back if not completely to racking their own servers partially. And so this type of can’t understand my bill, can’t figure out what to do about it. Issue is a problem for companies who spend a significant portion of their expenses on cloud, uh, services and also for the cloud industry in general.
And so your first step was, is we, we talked before was to say, you know, maybe what we need is better. Competition. Clients are locked into these cloud companies. Maybe they’re working with Amazon web services with AWS. They’re locked in and they can’t move out. What if we could allow them to spread their data across multiple clouds and then just shift from one to the other based on pricing.
You built it. How far, how did, how far did you get with that solution?
Leon: So , it was one of those ones that it was really hard to kind of agile your way into an MVP or at least. So I thought at the beginning, right. I felt like customers don’t have the imagin. In for this, we need to present the whole solution to them in order for them to get it. And I think that was a huge mistake and we had to pivot, um, and we probably should have pivoted sooner, but the feedback we got, we got feedback from a really large customer that said, Hey, Leon, we love your vision of how to arbitrage costs here, but we can’t even figure out one cloud and you’re throw.
Multiple clouds. Add us simultaneously all the cyber security implications. We can’t figure it out. So we can’t use your product.
Andrew: And so you didn’t want to go and pitch a client on this service until you had it fully built. You fully built it. You went to clients and they said, this is too much.
Leon: that’s right. And maybe it was a little bit of my arrogance. Like, you know, it’s kind of like that, uh, near miss fallacy. Like if you never get hit by a bullet, you think you’re invincible. Right. So I had gone through a bunch of startups. I’m like, Hey, I know what to build, what you don’t always know what to build, even if you’ve got a ton of experience.
So that’s why PI failing fast and, and getting feedback. You know, there’s a great book called, uh, a, um, uh, it’s, it’s just escaping my name, but it really describes like the agility of startup of startup life and,
Andrew: But in retrospect, if you, if you would’ve said, I’m gonna go talk to these clients beforehand and say, this is the problem that I had. Did you have, do you have a two great, here’s a solution that I I’m going to come up with. Would you wanna sign up for it? If you would’ve done that two step process, you would’ve known whether to build it or not.
Is that the right approach in retrospect?
Leon: maybe, maybe, and my thinking was, if we talk to customers about it, they definitely won’t get it. They, they, they won’t be able to. They won’t be able to reason about a solution that works this seamlessly, cuz we really had a, you could create an environment in our platform and you can create an environment in like under five minutes across two clouds with full mesh security.
And I just didn’t believe that customers would understand the simplicity unless we showed it to them, which it was a mistake. I think the approach that you described was probably the right approach in hindsight.
Andrew: How much did you spend going down this path before you realized it? And the money to
Leon: It’s probably a couple million bucks.
Andrew: start this business to start cast came from where. Okay.
Leon: We funded it originally, our co-founder team funded originally. And then we had a group of seed investors that had done well in previous enterprises, previous startups with us. And so I think we raised a total of $3 million, uh, for our kind of seed round. To build out a product to product market fit.
Andrew: Okay. So about a third of it, it seems like am I right? It went to this first version. Okay. And then how did you know what to pivot to? What were you gonna do?
Leon: Yeah. So we always knew that cost, uh, we had this perception that. Uh, cost was gonna be a big deal because we’ve seen it. We saw it in our own his lives and we said, okay, well, why don’t we strip out some of the complexity and what if we took the optimization engine that we built for multi-cloud right for multiple.
And we just focused it on a single environment. We don’t have to arbitrage across AWS and Google maybe, but what if we had better choices that we could make within a single cloud? And it turns out that was exactly what customers were looking. We had, uh, sorry, go ahead.
Andrew: I think what I didn’t understand before was the first version. Didn’t just allow people to spread their data and their, uh, processing power, but among multiple cloud vendors, it also understood where the right vendor was based on their usage. So that cast understood where to send the data, where to send the processing, where to send, where to have the work done.
Am I right?
Leon: Yeah,
Andrew: I didn’t realize even that first version had that. Okay. You did a lot of
Leon: benchmarked every computer and every cloud as a first step. So we went and we ran a set of tests on every single computer to figure out what it’s relative. Like you’re paying a dollar per hour. Let’s say, what does that dollar really buy you relative to what you can buy for a dollar over here. And that was a massive project data science project to try to understand.
Is everyone selling apples to apples or is it really more of an apples to orange situ?
Andrew: Okay. And so then when you did that, when you said, okay, we’re just gonna take everything we did instead of, instead of encouraging clients to use multiple cloud vendors to get full use of this product, we’re just going to make it work for one vendor. What happened?
Leon: so we, we rallied around an, an event called CubeCon, um, which is kind of the Kubernetes, everyone fly in. And this was obviously. Um, during COVID. So it was a virtual event. Um, everyone kind of rallies around cloud native technologies, which is fundamentally Kubernetes. And we should explain that for folks who don’t know what it is.
And we said, we are gonna launch just one cloud, which was Amazon. They have a service called elastic Kubernetes service, and let’s figure out how to make our product just seamlessly connect to that environ. Without us managing the environment, we’ll just connect in and do the optimizations. And we had a Eureka moment and the Eureka moment was.
What if we could, without any friction, tell customers how much they’re gonna save before they even implement the product. Before they talk to a human or a sales guy, they could just do it on their own and get a real report, a readout as to what things we would change in their environment if we were gonna optimize it.
And that was kind of the, the aha moment for us. And it became the aha moment for all of our customer.
Andrew: So you went into Cuban with this tool that if they connected it to their Amazon account, to their AWS account, you would tell them what you would do, how much they would save. And they could then decide, is this something that they’re interested in or not? And if they see in black and white savings that matter, they can call you and you have a direct lead for someone who’s interested.
Leon: Exactly right.
Andrew: Okay. That seems like a very logical sensical thing to do. I was talking to, uh, Patrick from profit. Well, he connects into Stripe and other payment processors. He said, you know what? We were just going to show bus. What they do now is just show businesses. What their churn rate is how many people are leaving and then say, by the way, if you like, we have an automated system that will automatically reach out to people who are churn and help them update their credit card information and save you money, and will only tell you when we save money for you.
But if you’re seeing in this dashboard, how much money you’re losing to churn, you wanna improve it. And we’re here to help. That’s the thing that I’m getting from both of you, the, the dashboard, the data that says here’s the pain. Now that you’re clear about it, understand we have a done for you service that will eliminate it.
Leon: and for me it was two things I want to give customers value. And this report was a way of giving them value without them plunking down their credit card or paying us anything. But for me under, it’s also a very important, um, competitive moat. So imagine we fast forward a couple of years and I’ve got this little agent installed.
Hundreds of thousands of Kubernetes, clusters, millions of them globally. It’s very different, difficult for anybody, including one of the big cloud providers to displace us as an authoritative source of truth, from a data perspective, like we’ve got the data and it’s just like, would you install two antivirus on your computer?
It doesn’t make sense. Right? You we’re already connected very hard to displace.
Andrew: plea on beyond the savings that you’re showing them. What else are you doing?
Leon: So that’s just the, the, the passive mode. Right. But when we, when we, um, actively engage, our
Andrew: Got it.
Leon: does a very different operation.
Andrew: Got it. And you’re saying once you’re engaged, once you’re actually optimizing your client’s usage of their cloud provider, that’s when you you’re, it’s hard to displace you, you, you’re not going to have two different peop two different companies do what cast does.
Leon: Yeah, but, but even in the first phase, it’s kind of, you know, my plan is to offer more and more free value. So that, like, imagine you were signed up to the free service and you would get a report every single week that would say, here are all of the teams using this cluster. Here’s how their cost breakdown here are the most wasteful teams here are, here are the teams that are not following cybersecurity, best practices.
Like you actually have actionable things you can do without paying a dime. And as long as we continue to add that value to the freemium product, I think that the traction in it will grow. and then it gives us the data. We need to train our algorithms.
Andrew: How do you know what’s wasteful? How do you know what a team’s usage mean is wasteful and what’s use useful.
Leon: There’s a lot of different. So we chose Kubernetes, which is a kind of a container what’s called a container orchestration platform. Cuz we have a couple of beliefs. So I should just tell you what those hypotheses are like, the way you package software today is through these things called Docker containers.
You may have heard of the term before. It’s just a very convenient way of packaging your software in a really tiny footprint. And then if that’s true, then you need a way to manage and deploy these Docker containers at scale. And that’s what Kubernetes does. That’s Kubernetes is a Greek word for. The project came from Google.
It was open sourced when it was at Google, it was called the Borg. They didn’t want to be so evil. So they renamed it to some, uh, nautically themed, uh, uh, product, but it is the kind of winning con container orchestration platform in the world. It is open source and the adoption is going through the roof.
So our contention is that in five to seven years, most enterprises will be running software this way. And rather than trying to optimize all different kinds of environments on any type of virtual machine, we just picked one and we picked one that was very compatible with kind of machine learning algorithms, because it’s a game like we think of it as a game.
Kubernetes is a game board with a bunch of rules and we can teach a computer to play that game.
Andrew: guess, I, I appreciate the way that you explained what you’ve said so far, but I’d still like to know a little bit about how can you tell if something’s useful? How can you tell if. Uh, waste wasting.
Leon: Yeah, so sure. I, I can get, give you a couple of details. So when you deploy one of these kinda yeah, like when you deploy one of these kind of, um, these components of your software, You’re usually deploying it on a computer instance that you’ve preselected. You said I’ve used these computers in the past and, um, I know exactly how they behave.
So I’m just gonna use these because historically I have a bias to using these well, it’s a market, it’s a market of hundreds of SKUs for in the case of AWS, you may not be picking the best computer. How do you know? Right. Like you haven’t benchmarked all of them. Like you don’t have you. Implicit bias. So, uh, one, the first thing our does is our engine chooses the lowest cost computer that is best for this workload.
And we match, we do this real time matching of workload to computer. That’s just one small and it’s cost driven as, as opposed to human driven.
Andrew: Okay.
Leon: And there’s a couple other examples I can give you. And there’s one really interesting kind of Freakonomics example. So let me explain, because this is an open market. There are some computer types that wildly fluctuate in price. So Amazon, for example, has this thing called a spot instance. It’s a spot computer.
And what that means is they say, look, we will give you this computer with no commitment for 90% off. You just pay 10% of the cost. But when a real customer comes around, we reserve the right to take it away from you within. Uh, two minutes notice. So customers only very rarely use those computers because it’s chaos.
How are you gonna manage the chaos? You don’t have a robot to manage it. So we do that for customers, but when we pick those computers again, remember Andrew, it’s an open market. So sometimes what happens is more powerful. Computers are cheaper. It’s an, or it’s a, it’s a paradox.
Andrew: Right.
Leon: You have a, a computer with extra bells and whistles on it that no one.
But because no one wants them, they’re overlooking them. And then the core price for basic compute is lower than the most popular computers. Well, our engine doesn’t care. We just choose the lowest price on the market at that moment.
Andrew: You move them, you move your client from the more expensive computer to the cheaper computer within minutes and back automatically.
Leon: 15 seconds. Every 15 seconds, we make a scaling decision on a, on the behalf of our customers.
Andrew: And that’s what the AI in your domain name is. It’s cast.ai because that artificial intelligence is what’s making the decision and moving things.
Leon: Yeah. And it’s a combination of algorithms. It’s not just one master model that, you know, plays the entire board. It’s many smaller models that work together to create a set of decision. It’s a decision tree, but they work all work together to create those decisions.
Andrew: All right. I wanna understand how you ended up getting more customers and how you decided to expand. But first I should, let me do it again. See if I could do it in three sentences or less. Here’s what send in blue does my sponsor. They do email marketing, SMS, marketing, chat, marketing inbox, uh, man, uh, already flubbed it.
You get to send email out, uh, to your customers. The beauty of it is that they combine it all in one packet. So your landing page talks to your CRM, which talks to the email that you send out, which then communicates back with, uh, your database so that you’re only sending the right message to the right person.
If you’re interested in signing up, I’ve got a discount for you@sendinblue.com slash Mixergy, send in blue.com/mixergy. You know, Leon, I have this theory that. That a lot of sponsorship messages are skipped or ignored because they go long and it would be even more powerful to go with one sentence advertising sometimes instead of one minute advertising, even if you’re saying less, as long as there’s a mix of both.
And I wanna experiment with that. And so, but I don’t think my sponsors want me to just go to one sentence since they’re paying. So I’m, I’m experimenting with three sentences or less on the spot. See if we could make it work. I don’t know. Do you get into this stuff? Are you, are you a marketer? It doesn’t seem like it, but you’re, you’re also in many ways, the voice of the company.
Leon: Yeah, no, I, I do really love to dive deep in marketing. Like customer acquisition cost is a huge, uh, it’s a hugely important metric for SaaS. And so we look at it all of the time. Like, how, how do we know when it’s time to invest in this, in the scale up of customer acquisition? So, and then I, I like your theory, Andrew, for me, if it’s the message is over before I can hit the ten second skip button on
Andrew: Right.
Leon: that’s a good sign.
Like I, I won’t bother. Right. Cause I have to get my phone, but if it’s a 32nd message, I don’t want to hear about vitamin water for, for 30 seconds. I’m gonna skip over.
Andrew: I’m gonna try more of that. I’m gonna see if I could consolidate it down to one sentence and then sprinkle that in every once in a while. So maybe some interviews will just have two ads, but only one second each, by the way, I’m on similar web, trying to get an understanding of where your traffic comes from.
I don’t think you’ve got the kind of traffic that. That can be analyzed on sites like similar web, because it seems like it’s, it’s not like massive amounts of traffic. It seems like it’s, um, there’s a lot of actually, I, I don’t really see it. Where are you getting your customers? How are you doing your marketing?
I shouldn’t say that it’s nothing. I mean, your, your global traffic is what in the half, a million a month rained. Right?
Leon: yeah. And, and for.
Andrew: that.
Leon: Yeah. And for a B2B, uh, SaaS platform, that is a, like, I I’ll tell you, we are out marketing all of our bigger competitors, those with much bigger budgets.
Andrew: Here’s what they’re showing me though. It’s search. So it’s a lot of content and then it’s direct traffic, right? And no, sorry you go ahead.
Leon: Yes. Yes. And part of it is, so there, sorry, go ahead. I was saying like, part of it is there’s a lot of gorilla, um, work being done in developer forms. We generate a lot of content that is useful for people just to read our posts. You know, I’m personally writing a lot of content. So my whole team is doing so as well, engineering content. And then in this environment, in this macro environment, it’s like, I have a lemonade stand in the desert, honestly, like we’re , this is not.
People are looking for savings right now. Everyone’s looking to sight in their belt and we have an obvious way of doing it.
Andrew: You know what I was wrong. I think I was just misreading similar web. I do see this. I see that you’re big on Reddit. For example, I’m assuming that you’re in those forums on a regular basis.
Leon: Yeah. That’s a portion of the gorilla tactics.
Andrew: uh, what, what else? So it’s you have someone in house who go and I’m assuming it’s all, it’s yours who are going out to the forums. What else is working for you for getting new customers?
Leon: uh, so it’s really important to get in front of analysts early, and that has a very long payoff cycle. So you might get in front of Gartner, you know, this year, you may only start to get them recommending you a year or two. So the best way to do that is to start early. The second way we do meetups in person and virtual.
So go Lang meetups, Kubernetes, meetups, other open source projects. Our team is passionate about that. We have a very strong support for women in tech and we sponsor a lot of those, uh, trade shows are extremely like Cuban and Valencia. This year was, I can’t say the foot traffic was off the charts, but every single person that came to the.
Was a whale in their own. Right. And we ended up with, you know, if we close 20 deals out of that show, it’s gonna be a home run.
Andrew: Why this what’s your goal with being on my podcast? I can’t imagine that you’re looking for customers. What are you trying to do?
Leon: So this industry is really complicated. Andrew, the cloud industry is really, and I’ve been doing this for a while like this, I, I guess I’m into industry for 25 years. Um, and I really feel there’s a big gap between. Current set of engineers are going through and all of the lessons learned and some of the hidden gotchas in the industry.
So I really want to kind of evangelize the best practices and we’re gonna start with cost because it’s a major pain point, but cybersecurity, disaster, recovery availability. These are all lessons that have been hard learned that I really want to share with the C.
Andrew: And it’s, it’s raising profile, getting, getting people to understand what’s possible and just planting seeds. It seems like it’s not about direct response.
Leon: it’s not a direct response strategy. No, we would. Uh, but it is about becoming, having our company, which is a very young company, become an authoritative figure in some of these discussions and. And there’s a lot of interesting ones coming up, whether clouds are behaving monopolistically or at least in an oligopoly.
So there’s, there are a lot of interesting issues. Some of them are existential for this business, and I think it’s our job to elevate them to, uh, average, uh, listeners and readers.
Andrew: What about, uh, how you understand your customer’s needs. Now we talked in the beginning about, maybe you could have talked to your customers before explained what you had in mind and then gauge response. Now that you’ve got that behind you. What do you do to understand what you should be adding what’s gonna work and what you should spend your resources on?
Leon: Yeah, that’s a good question. And, you know, we did a com kind of almost a complete 180 on the, the strategy there. So we have a. Uh, as part of our UX team. So UX stands for user experience. We have a research component that research component creates Figma. Basically we create like when we’re gonna build a new feature we’ll and if it’s UI driven, we’ll CR we’ll create a Figma mockup that’s interactive.
Before we even write a line of code now directly to customers, this is the way it’s gonna work. Give us your feedback. Not only do we talk to our existing customers, we have a service that we pay to bring us like companies that aren’t our customers and say, okay, you, you have this problem. You’re not our customer.
Tell us what do you think of this solution? And they give us really honest feedback. So we’ve done it, uh, almost a complete 180. And I would say that’s thanks to my team for pushing me to be more data driven versus instinct driven. Yeah,
Andrew: I’m looking at Riverside, this software that we’re using to connect and talk today. I love that. It tells me things like what your camera is. Apparently you’ve got the log 4k pro camera. The reason that I’m bringing this up. Oh, is that right? I just thought your eyes light up. The reason I bring it up is because I feel like you care about details like that.
Like how you’re coming across on camera, what’s behind you. Things are, are thought out with you. I wanna get a sense of how you are as like a person. What do you do to stay productive? What are the details you care about? What are the things that makes you that, that productive person that you are.
Leon: Yeah, I, I don’t think I’m super productive these days. Uh, and I struggle with that Andrew. So like, uh, for me, uh, order is very important. My desk is extremely clean if not empty. Uh, and it, if people who know me understand that’s an obsession, I can’t really start working unless every, like, are you one of those guys that makes your bed when you get outta morning?
Andrew: I always have. Teasy
Leon: Yeah. Like I, that, like
Andrew: yeah.
Leon: I have that very long routine that it takes me to actually get in front of the keyboard. And, you know, I I’ve suffered with attention deficit disorder for many, many years since I was a kid. And there was nothing back then. Like, you know, if you were fidgeting in your chair, when you were kid in school, they would just tell you to stop fidgeting.
It, it wasn’t. Okay. Right. And now thank goodness. We have a lot better appreciation for how to help kids. Through those type of things. So I really am highly focused on, uh, how to bring focus and flow to my work. And, um, it’s a combination of a bunch of things that I do physically, and some things that I do mentally, um, which we can talk about, but it, it is a struggle for me.
And it’s something that I work at every day.
Andrew: It seems like part of it is eliminating distractions. I’d love to get into it. Take me into Leon’s mind and his process. What do you.
Leon: so physically there’s four things that are important to me. Uh, extreme exercise, um, intermittent fasting, which I only started a few years ago and it’s been literally a lifesaver for me. Um, uh, hot, extreme heat and extreme cold. And I’ve only added those two things. In the last year. And I find that the combination of those is helping me, uh, recover more quickly and is helping me mentally.
And here’s the analogy I like to use. If you come off the mat, in my case, martial arts, and someone’s just been trying to kill you literally for an hour, it’s much easier to kind of take the stress of the rest of the world. Like these other problems don’t seem so big when you look at it like, oh, I was just about to be choked.
No, no, no. I can handle this sales problem or I can handle this engineering problem. So
Andrew: So it’s intermittent, fasting, hard workouts going from hot to cold. This is like the, the sauna to cold to an ice bath and back
Leon: I, I don’t usually do them. I do them separately. So like, like we do like a, a cold shower every morning, um, for at least two minutes. And it’s hard for you to do Andrew and Austin, cuz the water is worn in the pipes. It’s it’s not, you’re not gonna get a lot of cold water. You have to make a deliberate effort.
Um, and then sauna. Yeah. We installed a sauna during COVID to help with that portion of it.
Andrew: Okay. And then what else is it beyond that fasting, intermittent fasting your hard workouts, hot and cold.
Leon: And, and, and then, and then it’s a set of things that I help myself. With, uh, uh, mentally and emotionally, I don’t, um, I don’t meditate in the traditional sense that a lot of people recommend. Um, but I do read a lot of fiction and that fiction for me is the meditating piece. Like I really enjoy, um, reading fiction and that helps me completely disconnect from engineering work and technology.
Andrew: And then clean everything, no distractions. When you’re sitting down, you might spend 20 minutes clearing just to make sure that when you’re sitting down, there’s not a thing for you to fidget with, or just say, you know, I better pay this bill or look at this letter instead of doing the work.
Leon: Yeah, it drives my wife crazy. And I know you’re gonna think it’s ridiculous cuz I, I start by washing the dishes every morning and she hates it. Like she she’s like get out of my domain, don’t touch my stuff,
Andrew: oh, I would hate to do that.
Leon: Yeah, but it’s cathartic for me. Like, I feel like I got my cheaped up, uh, rush of dopamine by washing the dishes and now I can go on and do bigger and better things throughout
Andrew: Okay. All right. Well, you’ve definitely done that. The website for what’s the cast part it’s I was gonna say cast.ai. What’s the cast part. Mm.
Leon: It, it was originally an analogy around the nautical theme. So if you cast, uh, a web. You’re gonna catch more than one cloud. You’re going to catch more than one set of ways of optimizing your environment.
Andrew: Got it. cast.ai and I’ll thank the two sponsors. When you’re hiring developers, go to lemon.io/mixergy. And when you’re looking to do email marketing, SMS marketing landing pages of works go to send in blue.com/mixergy. Both, both URLs will get you great deals, Leon. Thanks so much for being on here.
Leon: Andrew. Great questions. Thanks for, uh, thanks for digging.
Andrew: Oh, man. Thanks. Bye everyone.