Why is this founder breaking things on purpose?

+ Add to

Sometimes entrepreneurs feel that if they take a real job they are failing their true self. I think that’s bullshit.

Joining me today is an entrepreneur who worked at Amazon and Netflix. He solved big problems for those companies. He took that experience and started a company to solve the same problems for companies that weren’t as big as Amazon and Netflix.

Kolton Andrus is the co-founder of Gremlin, an on-call Chaos Engineering software platform. We’ll talk about what that means in this interview.

Kolton Andrus

Kolton Andrus


Kolton Andrus is the co-founder of Gremlin, an on-call Chaos Engineering software platform.


Full Interview Transcript

Andrew: Hey there, freedom fighters. My name is Andrew Warner. I’m the founder of Mixergy where I interview entrepreneurs about how they’ve built their businesses and I do it for an audience of real entrepreneurs. And the thing about real entrepreneurs is often, they feel like, “If I ever get a job, I’m not a real entrepreneur. If I ever get a job, I’ve kind of betrayed myself and I’ve shown the rest of the world that I’m not really the entrepreneur that I said I was.” And I think that that’s just absolutely BS and you can kind of see that in the interviews as they’ve come up. And if you haven’t yet, you’re going to see it today.

Joining me today is an entrepreneur, Kolton Andrus, who worked at Amazon, who worked at Netflix, who solved big problems for both those companies and then said, “Hey, you know what? Maybe other companies that aren’t Amazon, aren’t Netflix, aren’t necessarily even as big as those companies, maybe they have the same issue that we did. And if I solved it for Amazon and then I solved it again for Netflix, I bet I could solve it for these other people too.”

And he started a company that he calls Gremlin. And if you kind of look over . . . No, we’re going to probably edit it out. Everywhere I look, I fricking see a gremlin over his shoulder in, like, the favicon, of his website, all over. So he’s clearly having fun with this name, Gremlin. He even owns gremlin.com. And when I asked him for, like, a professional one-sentence description of what Gremlin was he says, “Failure as a Service.” And I think he’s looking to create, “A what?” type of response, which he’s going to get from me right after I tell you that this interview is sponsored by two companies that do phenomenal work. The first will host your website right. It’s called HostGator. The second will help you hire your next great developer or designer or finance person. But first, Kolton, what do you mean by, “Failure as a Service”?

Kolton: Yeah. It is intended to get that, “What?” response. So I’m glad you keyed on that. We break things on purpose, but we do it with thoughtfulness, with . . . We’re there to mitigate risk. We’re breaking things to find weak spots and to make them better. The analogy I always draw is that of the vaccine. If you go back 200 years and you say, “I’m going to inject you with this disease. Is that a good idea?” You might have got mixed results. But what we’ve learned is that that scientific process of hormesis is very valuable. Well, the same thing applies in our tech and in our companies. By causing bits of failure, by causing bits of harm, we can uncover weaknesses and we can build immunity so that those issues aren’t issues in the future.

Andrew: And you do this for software companies. Give me an example, if you could.

Kolton: Yeah. Some of our customers, people like Twilio, people like Walmart, Under Armour, e-commerce, SaaS, financial companies, companies that care a lot about the cost of downtime. If my website is down, I’m losing hundreds of thousands of dollars in revenue. How do I prevent that? And, you know, the classic operations approach is after it’s broke, how quickly can I fix it? And the key difference with chaos engineering or resilience engineering is, “Hey, let’s go out before it breaks. Let’s find it. Let’s save that time and pain and make things better.”

Andrew: So what you try to do is create . . . Like, I think you were telling me before we started, disconnect to one of your clients from Amazon’s computer services to see what happens to their software and then when it breaks, come back to them and say, “Aha. You have a point of failure. You’re too dependent on them. Let’s find a solution, in case they go down.”

Kolton: Yeah, exactly. So about two years ago, Amazon S3 failed. And it took down a big portion of the internet. What happened is a lot of people discovered that they had very [carved 00:03:31] critical dependencies on Amazon S3. Many of them weren’t aware of that. What we would do is we’d go in and we’d help those teams safely and securely test the failure of Amazon, of Amazon S3. And from that, they’re going to learn two things. One, is this a critical dependency and two, can we make it a noncritical dependency? Can we make it so that if Amazon fails, our customers don’t see the pain? And, if at all possible, that’s the goal. Graceful degradation.

Andrew: I always wonder is it tough to convince a client to pay for something that they don’t feel the pain of until it’s too late?

Kolton: It is one of our hardships is if they had an outage in the past week or month, then they’re probably thinking about ways they can do better. If things fail all the time and they’re just used to it, maybe they just think that’s how business is run and life doesn’t get better. And so being able to quantify the cost of downtime, the loss of revenue, the loss of customer trust, being able to quantify the engineering effort. When an outage happens, we see the public-facing element. But engineers will lose dozens of hours and dozens of engineers digging into it, understanding it, trying to prevent it. That’s a lot of money that could have been invested in your product and in your company [inaudible 00:04:47].

Andrew: But you don’t feel it until it’s too late. So I won’t call them out, but there’s a company that does something similar. They are there to help with issues when the site goes down. One of the things that they will sometimes do is just send cupcakes to people whose sites go down saying, “We know how painful this must be.” And I don’t know what they do to make it sound soothing and not, like, a jerky thing, but it kind of opens the door to a future conversation. Do you do anything like that so that you are top-of-mind when people are in pain? And you’re smiling really big.

Kolton: Yeah.

Andrew: What do you do?

Kolton: We’ve debated this a lot because, on one hand, we don’t want to be ambulance chasers. We don’t want to come right after you’ve failed and say, “Hey, you’re dumb and you did things wrong. And we’re going to tell you how to do it better.” You’re not going to win friends that way and it doesn’t really work. But we’ve created these hugs-op boxes. And so the hugs-op box, it’s a little green box. It’s got gremlins on it. It’s got tea in it. It’s got a stress ball. It’s got some snacks. It’s got some things to help you because we’ve been engineers. We’ve been on call. We know it sucks. And we want to convey that same sentiment. It’s okay. Life can be hard. We want to help you when you’re ready.

Andrew: And you send this to them before there’s a problem, not during a problem.

Kolton: Well, so if I knew who was going to fail in advance, I would send it to all of them before it happened.

Andrew: But everyone’s going to fail in advance. So you’re sending it to them . . .

Kolton: You’re right. [inaudible 00:06:06].

Andrew: Right. I mean, everyone is going to fail, but you’re not sending it to them just to try to start a conversation. You’re saying when you have a problem, you send a hug-ops box. That’s like hug operations people. You send them a box of stuff. Got it. So it’s similar to the cupcakes that I mentioned earlier, a way of getting people to recognize that you’re there, if they need you for the next time. So, you know, we’ve got a producer who did a pre-interview with you. We also have researchers who looked up, like, what your revenue was. One estimate that we saw . . . I’ll tell you where, a site called Owler, was $2 million in revenue. Does that sound right?

Kolton: That’s a pretty close estimate.

Andrew: Pretty close. Okay. And the company was started, what, 2016 I think?

Kolton: Yeah, three-year anniversary in two weeks.

Andrew: And you’re a guy who . . . I talked about how your background was at Amazon, but let’s go back a little bit further behind. I wonder if you could tell us what Kolton Soft was going to be when you told your grandmother about this business. Do you remember that?

Kolton: I did. I remember having a conversation with my aunts when I was 10 years old. And I told them, “One day I’m going to start a software company.” Now I grew up as a gamer. So I always thought it was going to be a gaming company. And when I got to college, I found that gaming was a lot of hard work and not as much fun as it sounded and enterprise paid better. So I’ve always had in my mind I wanted to run a company. I wanted to build a company. I kept track of people I worked with in college and in my early career that I knew were smart that I’d want to go back to later.

Andrew: You literally, as you saw people in college, filed them away in the back of your mind as potential people to work with, to hire, to partner with?

Kolton: Yeah. And my co-founder is one of them and a bunch of the people on my team are some of them. I’ve gone back and used that list.

Andrew: Like what was it about your co-founder that back then made you say, “I think I could work with him in the future. Let’s file him away”?

Kolton: My co-founder, in particular, was about trust. We became friends. We worked closely together. We met at Amazon and we worked there. He helped me build this tooling. So I built a version of this for Amazon and then a version for Netflix. So one, you know, he understood it. He felt the pain. He helped me build it. But two, we became good friends. I could trust him. I could talk to him. I knew that we’d been through some good times and some bad times.

Andrew: In school?

Kolton: This was at Amazon.

Andrew: But in school, you kind of filed him away, did you, or you’re saying at Amazon?

Kolton: I hadn’t met my co-founder at school yet.

Andrew: Got it. Okay. We’re talking about Matthew. How do you pronounce his last name, Fornaciari?

Kolton: Fornaciari

Andrew: Fornaciari. Now the guy who was, like, so thinking ahead 10 years old telling his aunts, “I’m going to start my own Microsoft-type company later on.” So thinking ahead that even he realizes gaming is not where it’s at. “I’m going to start enterprise.” How does a guy like you then end up working at, and I talked about two different companies, Amazon and Netflix, but there have been others. How didn’t you just start a new company right out of school?

Kolton: I mean, if I had the right idea and knew what to build, sure, I would have. And I worked for a couple startups before Amazon and Netflix. And that was a great experience of, you know, I was the fifth engineer and the 30th employee at a startup in Salt Lake. That was great, great insight into how a business was run. But in my opinion, you know, there’s a lot of execution. There’s a lot of networking. But you have to solve a real pain. When I arrived at Amazon, what I saw was this very fun, but painful thing that . . .

Andrew: What was that?

Kolton: . . . people could do better at. Well, the fun part is the chaos engineering, the breaking things on purpose. You say that to an engineer, their eyes light up. They’re going, “What?”

Andrew: Why?

Kolton: Well, it’s kind of counterintuitive. Like, most of software engineering is, “Let’s write code. Let’s make sure things do the right thing. You know, let’s build and deploy.” The operational side has always been a second-class citizen. It’s never been as exciting. And everyone expects you to write good software and they expect it not to fail, which is also somewhat of a high bar, somewhat unrealistic. And so when you tell people, “Hey, we’re going to kind of throw that all out the window. For the next hour, we want you to think about what could go wrong.” It’s kind of the hacker mindset. “You know, hey, go in and break things and mess things up and find out where it’s weak.” And engineers love that. I loved it.

Andrew: You know, I do see bigger smiles on my kids’ faces when they, like, whack down somebody’s LEGO, even if it’s their own creation versus making it. And that’s what you’re kind of talking about. So that’s the fun part. You said though it didn’t matter until you found or you didn’t want to start a company until you found something really painful. What was it that was so painful about this issue when you got to Amazon?

Kolton: Yeah. So I was part of the retail website availability team and it was our job to make the website didn’t go down. And Amazon is one of the best companies in this space and I didn’t realize how good they were until I got to see much more of the industry. But even they don’t do it well, not as well as they could.

Andrew: You’re talking about 2009.

Kolton: Yeah, this is 2009, 2010.

Andrew: But Kolton, in 2009, I don’t remember Amazon going down on a regular basis.

Kolton: Do you want me to send you some “TechCrunch” articles?

Andrew: Is that right?

Kolton: I remember almost every one.

Andrew: I don’t remember them. So this was going on and my memory just has wiped them out because I don’t think of Amazon as the kind of company that goes down. It just kind of works. Got it. And so you go in there. You see the site goes down. You know exactly how much money people are losing. You realize, “This is an important place to be because first of all, it’s fun creating chaos to try to figure out the problem and number two, I can save this company measurable amounts of money and keep them from being in ‘TechCrunch’ for the wrong reasons.”

Kolton: Absolutely. Absolutely.

Andrew: And so you get in there. And did you create this product from scratch or how far along were they?

Kolton: Yeah. So somebody had come up with the general idea and pitched it in the Amazon one-pager. You know, “Here’s this thought for something we could build that . . . ” And it was a bit like Chaos Monkey. Chaos Monkey’s Netflix is an open source tool and it randomly reboots hosts. And so the initial pitch was, “Hey, we’re going to randomly break stuff and see if people figure it out.” And they said, “Here, Kolton. Go build this. Figure it out and do it.”

Andrew: Really? Okay.

Kolton: And so it was months of talking to teams. You know, our team already managed every incident and every post-mortem. So, you know, I read every failure that was documented at Amazon and that was a good way to understand what could happen. We had an internal process, call it game day, where we would go do this. We’d go pull cables and shut down racks and see what would happen. But what was missing there was if something went wrong, it was a massive impact. It could take down . . . You know, there was a day in Seattle where no engineer could do work because the internet was broken for the company. That was a hugely expensive experiment. And so we realized we could do it in a safer way that mitigated a lot of that risk.

So we built in things like undo buttons and rollbacks and deadman switches so that if things went wrong, the system wouldn’t cascade and fail. And that was a huge part of getting people to do it. In classic Amazon fashion, we built it as a service. So we built it with a user interface and an API and we said, “Engineers, come do the right thing. We’ll make it easy.” And we didn’t [deliberately 00:12:53] hand-hold.

Andrew: What do you mean by, “Engineers, come do the right thing”?

Kolton: In general, if you want engineers to do the right thing, you need to make it easy. And at Amazon, especially in 2009, 2010, that was, “I expect to hit a website, an internal website, that’s going to be the user interface for this tool. I expect there to be an API so that if I want to automate this type of activity or I want to run it in the background, I can do that. And, you know, if I need help, I want to come to your team and I want your team to give me guidance. But if I know what I’m doing, just stay out of my way. I’ll go use your website. I’ll go get the value. And I’ll get on with my job.”

Andrew: So I see. You weren’t creating this as a service that others outside of Amazon could use. But you were seeing the engineers inside of Amazon as your customers and you knew that they were not going to implement this thing that could disrupt their lives if it was also going to be hard to implement, hard to use, hard to make sense of. Got it. And so you were thinking of them as customers already thinking of this as almost a product. And would you literally take down Amazon for even a moment while you were testing something and then realizing there was a problem?

Kolton: So I’ve never caused a public-facing outage as part of my failure testing. And one of the ways that we mitigate that risk is this concept of the blast radius. Whenever we run an experiment, we want to run the smallest experiment that will teach us something. So if we can break a single user and learn something, if we can break a single host or a very small percentage of traffic, 0.01% of traffic, if that uncovers a bug, that’s all the risk we need to take and we can fix it and find it. Now, later on, we want to scale up and we want to test it at large scale. So a common pattern was 0.01% of traffic, 1% of traffic, 10% of traffic, 50% of traffic, 100% of traffic. And we only got to that point if all of the earlier steps worked well. So that was the way that we didn’t break Amazon publicly.

Andrew: I was looking to see when Chaos Monkey, Netflix’s product, was started. It’s 2011. This is after you started at Amazon. Was there a history of this existing? What did you base this on?

Kolton: Thanks for stroking my ego a little bit. I get to tell the story that before anyone had ever heard of Chaos Monkey, not only did we think about this, but we had built a fuller, more sophisticated platform at Amazon to do it. Chaos Monkey only reboots hosts. The platform we built and the platform we have at Gremlin is a dozen different failure modes. You know, discs fail, networks fail, time fails, processes fail. But that was part of why I went to Netflix. When I got done at Amazon, I thought, “Oh, you know, I really enjoy this. I want to go deeper. Who knows the most about this? Netflix.” And I went and joined Netflix.

Andrew: Do you think that Netflix got more attention for this because they named it? Is this, like, a marketing thing that Seth Godin would say, “Whoever names it gets to control it or gets the credit,” or what is it?

Kolton: Yeah. I mean, so Amazon’s MO is that engineers don’t talk publicly about what they do.

Andrew: Got it.

Kolton: You know, they viewed this as a competitive advantage. They wanted to keep it to themselves or they weren’t comfortable . . . You know, in general, they’re not comfortable letting their engineers talk publicly. Netflix takes the opposite approach. Blog about it. Give a conference, talk about it. And their culture is more accepting of that.

Andrew: You said, “When I was done at Amazon . . . ” What do you mean by, “Done at Amazon?”

Kolton: Well, the other joke I like to tell at Amazon, I was there four years and a day. It’s the four-year vesting cycle. I did my tour of duty. I built Gremlin. I built the failure testing tools. I spent a bunch of time espousing that and helping it drive adoption. We had hundreds of teams use that out at Amazon. And then I managed . . . I worked on performance and I hit my four years and I said, “You know what? Middle management at Amazon is tough.” It taught me a lot. It prepared for being CEO. But I was ready to go. I wanted to go deeper technically. And so that’s why I was looking around and Netflix seemed like a great choice.

Andrew: All right. I’ll talk about my first sponsor and then get into what happened at Netflix and how that led to Gremlin. My first sponsor is a company called Toptal. Kolton, do you know them yet?

Kolton: I don’t.

Andrew: Cool. I’m about to introduce you to them. In fact, usually, I talk about their development side. Today I want to talk a little bit about their design. There’s a company called WorkRamp that provides online training solutions for companies like PayPal and Zoom. Zoom is the software that you and I are using. They created this really highly functional product and they wanted to improve the user experience and the design to allow them to get more and further into the enterprise space. The problem was that they couldn’t get an internal designer who could get the feedback and design direction that they were using implemented.

And they said, “You know what? Let’s just go to Toptal.” With Toptal, they were able to work with a designer who fit seamlessly into their team as if that person was actually on their team. They were able to get started really fast. And as a result of implementing it, not only did sales for the product increase, but it also allowed them to get funding from companies like Y Combinator, from who else, or investment firms like Y Combinator, from Slack, from Alexis Ohanian known for Reddit, and Joe Montana. What the hell is Joe Montana doing investing? Everyone I guess is now investing. Did you see Snoop Dogg with, like, a bowtie out in, like, the stock market?

Kolton: It’s crazy. It’s crazy.

Andrew: Yeah, everyone’s now in tech. Startup is the place where it’s at. So I guess that’s why Joe Montana’s in it. Look, guys. If you’re out there and you’re looking to hire developers, designers, or even finance specialists, I really urge you to go to Toptal. As soon as you go to this URL I’m about to give you, press a single button. Once you press that button, you can talk to somebody at Toptal. Tell them about your development needs, your design needs, whatever. They will tell you if they’ve got someone internally that you can just get started with, often within days, or they’ll tell you what they’ve told some people who have done this. “Sorry, it’s just not a good fit.”

They’ll be open with you. Don’t complain to me when it’s not a good fit. I’m glad that they don’t take your money, if it’s not a good fit. Here’s where you go to hire really the best of the best. Toptal.com/mixergy. Top, as in top of your head, tal, as in talent, .com/M-I-X-E-R-G-Y. No wonder Andreesen Horowitz invested in them. Great company. I throw around all these names like Alexis Ohanian, Andreesen Horowitz, right?

Netflix, you went in. What did you do that was different at Netflix? Did you, like, just go out because you’ve got one of those . . . ? It did?

Kolton: Yeah, it totally just, like, tripped a motion sensor and I’m going to stand up and make a fool of myself and . . .

Andrew: Is that one of your decisions internally to say, “Hey, you know what? We have to control costs at every level,” or is that just something that happened to be there?

Kolton: It’s funny you mention that and this just happened to be here. But growing up at Amazon, there’s some Amazon core values that I identify a lot with. Vocally self-critical, customer focused, bias for action, but frugality is a very important one. And ensuring that your company is being thoughtful about generating revenue and managing expenses, we’re not the startup with, like, you know, cool tables and, you know, free lunches and all of these things. We’ve taken a very pragmatic approach. We take good care of our employees, but we do it in a way that, you know, is synergistic with the business.

Andrew: What’s an example of something that you wouldn’t spend money on but Silicon Valley companies just take for granted or something that you’re kind of proud that you don’t do? What’s your door on two pieces of whatever those things are? Yes.

Kolton: The door desk. Oh man, I’ve cut myself on those. They’re dangerous at times.

Andrew: You literally had one of those door desks.

Kolton: Oh yeah.

Andrew: At Amazon.

Kolton: Absolutely. Absolutely.

Andrew: And you really cut yourself.

Kolton: Well, because the joists are just these metal kind of sharp jagged pieces. And if you catch them wrong, they can be dangerous.

Andrew: Wow.

Kolton: I would say free lunches is the one. That’s pretty much status quo in Silicon Valley. I read a great book, “Work Rules!,” by Laszlo and he used to be head of HR at Google. And he spoke about how, after a time, people just really didn’t appreciate it. They just came to expect it. And so, you know, will I take my team out to lunch? You know, we’re remote by default. I’m only in my office a day or two a week. When I’m here, I take people out. I take care of them. But do I have lunch every day? No. No. I think there’s better ways to make sure my team is happy.

Andrew: All right. So Netflix, let’s just talk about when you got there, how similar was it to what you did before at Amazon?

Kolton: What’s funny is when I first joined Netflix, I didn’t join the chaos team or the resilience team. I joined the team that managed the API. And we were just a critical part of the Netflix experience. If we went down, Netflix was down. Twitter was on fire. And I wasn’t actually focused on chaos engineering when I arrived. And three months in, I was like, “Hey, do you know what you guys need more of? This chaos engineering idea.”

And they were using Chaos Monkey, but, again, Chaos Monkey randomly rebooting hosts, you hit the limit of that approach fairly quickly. And so what I had was an opportunity to build a new system and I built a new platform. It was at the application level. The crux was it was very precise. That ability to break a single user or a single device or 0.01% of traffic, that’s where we really honed that. And we built that platform for the whole company as a service, a good user interface, a good API. We taught some people the first time how to do it and then teams picked it up and did it on their own. And my claim to fame there . . . When I joined, we were around three nines of uptime, give or take. When I left, we were at four. So we saw a noticeable difference.

Andrew: You mean 99.9999.

Kolton: No, no, 99.9 is three nines, 99.99 is four.

Andrew: Got it.

Kolton: And this is another reason people need to be more comfortable sharing their metrics. Most companies aren’t above three nines. Most companies you interact with on a daily basis, even the really big names, have about eight hours or more of downtime.

Andrew: So I’m looking at Slack now as I do this [inaudible 00:22:42].

Kolton: And an hour is very difficult.

Andrew: Let me look at Slack. I don’t even know what search to look for to come up with this, but you know what? Some Android users are not receiving notifications. Slack right now has an alert on their site. All I could find was status.slack.com. I thought maybe they would be someone with a number I could see. Oh, wait. Can I do history? No, I can’t. All right. So you get in there. It was FIT that you created, right?

Kolton: Mm-hmm.

Andrew: Failure injection service testing. That was your thing.

Kolton: Yep.

Andrew: And so at what point did you decide, “I’m going to go out on my own and bring this to other businesses?”

Kolton: Yeah. So after Amazon, I thought, “That was a cool project. We provided a lot of value. Maybe that’s just something Amazon needs. Maybe everyone else doesn’t need it. Who knows if people will pay for it?” You know, I met one of my VCs at a party in between the two companies. And I gave him the quick pitch and that was all that happened. Nothing really went beyond that. But after seeing that Netflix, a company that was good at this and was known for it, needed better tooling and when that tooling and the approach was done well, saw a very large benefit. We didn’t have a good measurable benefit. But one of the executives said, “This is worth $50 million,” to us, that extra nine of availability. So there’s a huge benefit to the business there.

Andrew: Wait.

Kolton: That’s the point that I realized.

Andrew: Who said this? You said one of your investors?

Kolton: One of the executives at Netflix made the comment that when we moved from three nines to four nines, that that was worth $50 million to the business.

Andrew: Why?

Kolton: Customer retention, the customer experience. Netflix cares a lot about winning moments of truth. People have a lot of options for what they do in their spare time. If you go to watch Netflix and it fails, maybe you pull up Hulu or HBO. Maybe you go outside. Maybe you read a book. So they want to ensure that the customers have the best experience possible. And they recognize that it being fast and always up and working is a key part of that.

Andrew: Wow. I wouldn’t have that, “You know what? If Netflix is down for a moment, I play with my Wi-Fi router, I click around, and then maybe I come back to it later in the day.” But you’re saying they see measurable value in keeping their site up. Even for that tiny bit of time, it’s too painful. By the way, we were talking about Slack. Slack is three nines. It’s 99.95% uptime in the last 12 months. No, that was November. That was November. Actually, in a lot of months, they’re 99.95%.

Kolton: That’s not bad. I think for the year, I know . . . I mean, I don’t want to pick on Slack. You know, I use Slack. We use it for collaboration. I can think of three memorable Slack outages this past year that my team felt.

Andrew: No, I could see quite a few actually. They don’t do the year average, but I could see, for example, there was one month where they were 99.82% uptime. So right.

Kolton: Those are what kill your averages. Like, one bad month, if you have . . . Because again, if you want to get to four nines, you get less than an hour a year. And so if somebody has one month where they’re down for an hour, you’ve blown four nines for the year.

Andrew: So let me tell you something though. The thing that I would think, if you came to me and said, “Look. I’m thinking of creating this because even a little bit, even losing one nine is dramatic because look at what happened at Netflix and they know how many millions of dollars they’d lose, $50 million because of this, and look at what happens at Amazon,” I would say, “Those two companies are maniacs. They care about every little thing and no wonder they’re at the top of their game, but most companies don’t. They let that stuff go or they don’t think it’s as critical as say, employee lunches. I don’t think you’re going to be able to get customers.” How did you know that you would be able to get customers who cared about this and happened not to use desks that would give splinters to their employees?

Kolton: Well, I mean, you’re getting into the unknown aspects of being an entrepreneur, right? The answer is I didn’t know. I had a good gut feeling. I had two data points. There’s this trend in the industry. Everyone wants to be like Amazon or Netflix, whether they say it out loud or not. And they’re copying a lot of their architectures and a lot of their approaches. Everyone’s moved into the cloud and joined microservices. So everyone is going to inherit Amazon’s and Netflix’s problems.

As you get into these very complicated systems, it becomes harder to operate, not easier. Microservices doesn’t make your life easier. It makes it harder. And so people are going to need this or their software is going to break and they’re going to go out of business. So I think that competitive advantage that I saw I felt other companies would want. I did some exercises. You know, I did some cost of downtime, back of the envelope math. You know, what is a minute of downtime for Amazon? What’s a minute of downtime for Expedia or Twilio or Warby Parker or [inaudible 00:27:14]?

Andrew: Just dividing their annual revenue by minutes in the year, that kind of thing. So then I thought that you were going to tell me about this dozen companies that you talked to that you did your diligence process on and that once you talked to them they told you, “Hey, if you could actually create this, we’d love it.” But it doesn’t seem like it was that cut and dry.

Kolton: I mean, that was . . . So some of this is my mental model. You know, am I willing to take the risk? I have five kids and getting investment before I took that leap was an important part financially. When I met with the investors I liked, then it was, “Okay. Go talk to 10 customers and let’s see what they say. And then we’d be willing to buy it.”

Andrew: This is Amplify.

Kolton: This is Amplify.

Andrew: So one of the things that Amplify told you was, “Stop saving your money at Netflix. You could raise money today.” You’re smiling because it’s right or because I’m oversimplifying?

Kolton: No. There’s a fun story here.

Andrew: Tell me.

Kolton: The way that I met the Amplify team, I was in the lobby of the QCon conference. And I had just given a talk on what I had built at Netflix. And this guy comes up to me and we start having a conversation. And he’s asking me, “Well, are you going to do a company around this? Are you doing to take funding?” And I said, “You know, I don’t know if I’m going to take funding. I kind of want to bootstrap. You know, I don’t want to give up control. I want to make sure that things are done well.”

And he said, “Hey, I’ve got some counterpoints for you. You know, do you want to max out your credit cards and personally put your finances at risk to try this out or are you willing to give up some ownership and pay yourself a salary?” And I live in California in the Bay Area and it’s expensive. And so, you know, if I was going to bootstrap, it would have been another five years of me saving Netflix money and then putting it all on the line versus, “Oh, hey, now that makes sense. All right. Let’s talk turkey.”

Andrew: Do you live in San Francisco or just work here?

Kolton: I live in San Jose, South Bay.

Andrew: They actually have facilities for kids. Do they actually have kids there? San Francisco has zero kids over seven.

Kolton: I joke that you find a lot of the, like, families in the South Bay and a lot of the more single or earlier in their careers and earlier in their lives people . . . Not their lives, but earlier, less children in San Francisco.

Andrew: Yeah. You know what? Nir Eyal, when he was living in . . . He lived from Palo Alto to San Francisco. This would be the author of the book, “Hooked.” He said, “I’ve got a seven-year-old. Do you know anyone with seven-year-old kids?” I said, “Yeah, absolutely. I know everybody.” I swear, I looked around. Even at my kid’s school, I could not find a seven-year-old in this fricking city. I know they exist. I’m not pretending they don’t, but I don’t know where.

Yeah, okay. That make sense? So you said, “All right. We’re going to take funding.” They pushed you to expand your view beyond the back of the envelope to go to talk to potential customers. How did you put this team of . . . I don’t know that I should call them a team, but how did you get this group of 12 companies together?

Kolton: Yeah. You know, I had help from my investors there. You know, they had been talking to people. You know, as part of the due diligence process that investors go through, they start floating the ideas you give them by customers and executives that they know. And they get a feel for, “You know, does that get them excited? Is that a real pain point? Is it something they’d pay for?”

And then the next step is, like, you know, it’s basically me going and giving this half an hour or hour pitch. “Here’s what I did. Here’s what I’m going to build. You know, would you buy this?” And they back channel with the investors and they let them know, “Hey, you know, Kolton seems like he knows what he’s talking about. I would trust him. This seems like a valuable thing.” Some of our early customers were building this themselves and already had experience. And so there was some market validation there, as well.

Andrew: In the sense that they would use you, if you built something new instead.

Kolton: Yeah. It’s funny because out of those original 10, I don’t know if any of those are customers today. We have a lot of customers and we’ve had a lot of success. They felt like it would and some were banks that are slow-moving and may come around, but that’s the game.

Andrew: All right. And so what did you learn from talking to those companies that you didn’t know on your own beyond validating that this makes sense?

Kolton: Yeah. I don’t know if I recall any salient points from those first few conversations, other than trying to understand where the industry was at and knowing that it was early days. A lot of people said, “Ah. That sounds like a really good idea. I’d buy into it philosophically. I don’t know if we’re ready. I don’t know if we’re going to invest in it.” And so creating that urgency, giving people over the hump, helping them really get into it and get the value was something that I realized, you know, and I’ve only continued to realize this as the company’s gone on, Amazon and Netflix I felt, like, were slightly above the rest of the industry. And having worked there, like, the sausage is made the same everywhere.

Everyone deifies Netflix and they’re a great company, but they have bad employees and bad engineers. And they have failures and things go wrong. And so I didn’t think there was that much of a gap. As I talked to more and more companies, I realized that there was a noticeable gap. And really, companies are going to need to do this to mature and companies need to mature to get to where they need to operate their businesses, but they weren’t there yet.

Andrew: So the first client you got, the first paying customer, was . . .

Kolton: Twilio.

Andrew: How did you get Twilio as a customer?

Kolton: Oh, Twilio’s an example of kind of the friends and family approach. So the 12 I pitched, none were friends and family, but I had a co-worker, Bruce Wong. He used to run the chaos team at Netflix. He did some of the original Chaos Monkey blog posts. He had a lot of success. He joined Twilio. He spearheaded a lot of their platforms teams, their monitoring teams and he felt like . . . I pinged him one day on LinkedIn. I was like, “Hey, you know, we’ve got this beta. We’ve got this product. We’re trying to sell it. You get it. Can I come pitch it to you?”

And Bruce is a friendly, you know, but he wasn’t like my best friend. He wasn’t just going to give me a deal. It was like, “Well, come in and pitch us and talk to my team. Oh, we built something like this internally. Why is yours better? What’s your road map? What are you going to do? How are you going to price it?” And at the end of that discussion, they saw value. We ran some of these game days with some of their teams. Their video team is one of the best teams out there doing chaos engineering. A shout out to Tyler Wells because he does it well. Their service is, like, five or six nines. And a big part of it is because they do this kind of thoughtful proactive testing.

Andrew: So then what were you able to do that they couldn’t do? I mean, you, with a young company that was months old?

Kolton: Well, our product had a wider scope than theirs. So they had focused on a couple of failure modes. We had picked, you know, basically, everything that can go wrong in the infrastructure world. It was a command line tool. It was, like, XML-type configuration-based, I believe. We had a nice user interface. We had a nice API. So the crux of the argument is, “Hey, your engineers, to run these experiments, are going to spend an hour or two. And engineering time is very expensive.” You know, if instead, they could spend five minutes and answer this question, you know, what’s the delta there?

Andrew: All right. I want to talk about my second sponsor. It’s a company called HostGator. Kolton, I run. I love running. I’m going to run a marathon on every continent. I actually discovered this last year, that I had this goal 10 years ago. My friend, Jon Bischke, now runs Entelo. He forced his friends to sit down and write their one-year goal and then their 10-year goal. That 10-year goal seemed so insignificant. I said, “All right. I’m doing it because it’s Jon. He’s my friend, but I don’t want to give her . . . It’s too far.” And then last year, I found that goal and I said, “How did I not even hit this?” So I decided I’m going to run a marathon on every continent. And one of the first things that I did was I went to HostGator and I just created a website.

It’s runwithandrew.com. There’s hardly anything on there right now, but I’ll tell you what it gives me. It gives me something to just show people. So there’s a guy, my friend Devon, who has just been sitting at the office from time to time. When he needs a place in San Francisco I say, “Come. Use the Mixergy office.” He says, “What are you up to?” I said, “I’m going to be running this marathon on every continent. And here’s a website, runwithandrew.com.” And he looked at it and he goes, “Do you need someone to, like, help shoot video?” I go, “Yeah.” He goes, “I still have to run my company.” I go, “All right. We can do it on the road. Come shoot video.” So now we’re going. The thing is it has a life of its own, this idea of, “Are you going to run with Andrew right now or is there something going?

Kolton: Yeah. I’m checking it out.

Andrew: You can see it’s just a default, like, theme with a little bit of text on there. I’m going to run a marathon on every single continent. We’re going to document it on that site and having something to show you, look at how you just went there. If it was just an idea, it would feel like, “All right. Something that Andrew’s, like, going to do. Fine.” But the fact that you go to the website makes it feel more real. The fact that I could show it to Devon makes him have something more real.

Here’s why I’m bringing this up in the context of HostGator. Yes, I use HostGator, but the bigger point is this. When you take an idea and you give it a website, you give it life. You now have something that shows, “This is significant. This is big.” And it’s as simple as just going to HostGator and firing up a WordPress site. All right. So if you’re out there and you want one of your ideas to stand out, like, I could imagine, Kolton, you might decide, “We’re going to do a conference around this, like chaos conference,” or something like this. Does that exist yet?

Kolton: It does. We ran it in September. It sold out. It was great.

Andrew: Oh, you have one? It’s yours?

Kolton: Yeah, Chaos Conf.

Andrew: Does it have its own website?

Kolton: It does.

Andrew: It does. There you go, right? Now it could have been on . . . Gremlin’s a great domain, gremlin.com/chaos or gremlin.com/chaosconf. It could have been any one of those things. But the fact that you give it its own website means now people have something to show their bosses at work. Now people have something to show their wives, something that they’ll remember that they can take a look at, something that their husbands can say, “All right. I think this makes sense. Go on and take a week away from the kids and I’ll go and take care of this.” This is what I’m talking about.

And I just flipped this pen out of excitement. If you have an idea and you want to bring it to life, go to hostgator.com/mixergy. They’ll give you a super low price, make it easy for you to get started. And if you hate your hosting company, go to hostgator.com/mixergy too. They will make it easy for you to migrate. Of course, when you add that /mixergy, at the end, which, thank you very much. It means a lot to me. Number two, you’re going to get the lowest possible price as far as I know for HostGator. And number three, you’ll be tagged as one of my people, which means if you ever have an issue with any one of our sponsors, we’ll stand behind you and make sure that you’re taken care of the best of our abilities. I obviously don’t run HostGator, but I do have the ability to run my mouth with anyone. So I’ll take care of you guys. Hostgator.com/mixergy.

Kolton: So talking about those 5 and 10-year plans, you asked me, you know, about foresight to be a founder. Well, I’m a big fan of the 1-year, the 5-year, and the 10-year plan. And when I was in college and my 10-year plan was to start a company, it felt pretty far off. It felt like, “I don’t know. Maybe it would be cool,” but that’s how you get there, incremental steps.

Andrew: What do you do to find it? For me, I was just going through Evernote looking for something. And Evernote’s search is kind of wonky sometimes. And it kicked this up. And it was the best because it hit me in the face and I was too embarrassed to look at it, but it stuck in the back of my mind. What do you do to make sure you look at your 10-year goals?

Kolton: I write them down. There’s a lot of value in writing them down. I don’t know that I ever go back and look at them. I have all these notebooks and every now and then, I’ll peruse through them, browse through them. But one of the things I do is you just tell people. And me, sometimes I might be kind of shooting off my mouth. I’ve had a beer or two. I’m going to start a company someday. And honestly, again, for better or worse, I feed off of, you know, if some people chuckle at you or kind of scoff at you, that’s just fuel for the fire. I just put it inside and I’m like, “Good. You know what? All of my buddies in college, did they think Kolton was going to start a company? Maybe. Maybe not. But now do they? I hope so.”

Andrew: And so if you get a voice in the back of your head of one of your friends who says, “Kolton’s never going to do this. He’s just got too many family obligations. He’s always going to be, like, the nice guy who goes to work,” what does that trigger in your head? For some people, it would trigger, “This guy’s so right. I can’t do it.” What does it trigger in your head that’s different?

Kolton: Well, I’m a rebel. You know, probably not appropriate for this podcast, but . . .

Andrew: Go ahead.

Kolton: I can tell the story about how I got kicked out of high school, how I was homeless before I went to college, about some trouble I got in.

Andrew: How? No, you know what? This is totally applicable because not only did you not tell our producer, but the researchers found that you were a Boy Scout Master in Troop 317. I go, “What the hell? I don’t have anything like roughen [inaudible 00:39:00] to bring up.” So I just skipped over it. What do you mean you got kicked out? Bring up the stuff that’s not appropriate for us to hear.

Kolton: Well, first of all, I grew up . . . I was early on the internet. So I’ve been thoughtful about what I post online for a long time. And I’ll write a book one day and probably under a pen name. And I’ll give all the gory details. My story of getting kicked out of high school isn’t as fun as it sounds. I was a rebel and I didn’t go to class, mostly because I didn’t need to. And certainly that factored into it, but I was in marching bands. And our marching band was better than our football team. We won state. Our football team never did, but my senior . . . I needed to take a health credit to graduate. And I couldn’t take it my freshman year because marching band got in the way.

So when I got to be a senior, I had to take a packet course. And like any high schooler, I waited till the last minute. I did it on a Thursday night. It wasn’t done the Friday it was due. So I called in sick. I handed it in on Monday. And I came in and they were like, “You know what, Kolton? We’re not going to accept this. And you’re not going to graduate. And you need to go take the GED.” And I fought it for a while. And I talked to some people and the principal. And in essence, you know, as somebody who, you know, got good grades but didn’t go to class, they had no sympathy. So when I make a big contribution to the University of Utah, Sky View High School isn’t going to get a check because they kicked me out.

Andrew: I have that feeling towards NYU. They wouldn’t give me a dollar break on the price. I literally went . . . From the money from my dad’s cash register at the store on Myrtle Avenue, I would take the cash, go into school and pay them. They wouldn’t give me a single dollar off, and now they’re asking me for all this money. I’d love for them to come and ask me for money just so I can say, “Whatever you gave me, I’m going to give you 100 times that.” That was my dream when I was in college. What was the other thing where you got homeless?

Kolton: That story is probably not appropriate for this.

Andrew: Sure it is. Why not? Chaos Monkey, man.

Kolton: It gets in the details, but the short answer is after getting kicked out of high school, I met some kids that were traveling the country. They were kind of a homeless group. They were down from Seattle. We met up with them. I went up with them back to Seattle. His dad lived on this beautiful island, Vashon Island, in the Sound. We ended up on that island for the summer.

And what was funny is me and my best friend, Chris, we got to be, like, the adults. They let us sleep in the house. We drank beer with the old man. We were part of the crew. And all the other homeless riffraff had to be out, you know, out sleeping in tents on the yard. But yeah, we had three, four months of, you know, spent some nights on the streets of Seattle, went back and forth to Portland. We stayed in this old abandoned library on Vashon. I went back and visited it when I moved to Seattle. It’s still there. It’s still condemned. Yeah. And then eventually, I went back home and got married and got serious and went to school.

Andrew: And had five kids.

Kolton: But yeah, there’s still more color there. Hit me in person. I’ll give you the real version of this.

Andrew: I’d love it. I didn’t even know if you drank beer. I wonder if Adam . . . This is kind of unusual. We have Adam, his communications guy, listening in on this on Zoom. I wonder if he’s going, “We had structure. We even had, like, a phrase to explain what Gremlin is. Now he’s going off and telling people about how homeless he is.” How does it feel, by the way? I don’t know what he’s doing right now, but how does it feel to have Adam listening in right now, to have a team that when I was looking at the size of your company, I think Owler said 36 people you guys have internally.

Kolton: Fifty.

Andrew: Fifty now.

Kolton: Fifty.

Andrew: Fifty and CEO approval rating, according to Owler, is, like 94 out of 100. But what’s it like to get to 50 people so fast and manage so many people? How has that been?

Kolton: It’s been a hell of a ride and 2018 was a crash course in being an entrepreneur, for sure. The first year or two of the company is a lot like being on a team. You know, there’s a handful of people. You’re coordinating. You’re getting work done. I was marketing. I was sales. I was everything. And I was helping, you know, code and write the product. You know, 10 is a nice spot where we started to have, like, a person in every department, but then we grew teams underneath them and a ton of recruiting, you know, a ton of finding the right people. We’ve been lucky. I’ve got an amazing team. They’re very strong. We’ve had some people that were willing to take bets on us and join us. And we try to treat them well and take care of them. And the rest is on the books, I guess.

Andrew: You know what? I read an “Inc.” article about you. I was trying to get a sense of your management style. I’m sorry, an “Inc.” article about you where you said that you wanted to do HR, payroll, legal, sales, marketing, every part of it to understand what you were handing off to other people. Did I understand that right? That was you.

Kolton: Yep.

Andrew: And so when you took on HR, HR is a pain in the butt. Legal is a whole, like, world. People go to school for that. What were you able to do that gave you enough of an understanding without getting diverted from leading the company?

Kolton: Well, you want to be efficient with your time, but, you know, understanding how payroll works and going through that process and taxes and setting it up, understanding 401(k) and benefits. We give a 401(k), by the way, which is a little abnormal for a startup and something I’m proud of. Sales, in particular, one, I am or for some time will be the best salesperson because I’ve lived it. I’ve breathed it. I’ve felt the pain. I know the details. But by going and talking to those first 10 prospects and then the first 10 customers we closed, I really refined the messaging, the value proposition. I got to hear directly from my customers their pain and what they cared about.

The same with marketing. You know, we take a different approach to marketing. It’s not pitching. It’s teaching and educating and helping provide value. And that’s what resonates with engineers. That’s what resonated with me as an engineer. And so having felt all this pain and done it myself when I hired people, I could give them that guidance. And when they asked me questions, I had context, I knew the tradeoffs, and I could hold them accountable. I knew if they were . . . You know, if I can do something and they weren’t able to do it, you know, as an engineer in sales or as an engineer in marketing, I’d step back and say, “Well, I figured this out. You know, what’s different here? What am I missing?”

Andrew: So you know what? The one negative feedback I gave [Arian 00:45:20], she’s the producer who talked to you before we started, right, or was it Brian? It was Arie.

Kolton: I think it was Arie.

Andrew: So the one negative feedback I gave her was, “You should have pressed him even harder on how he got the next batch of customers. What we found was yes, he got Twilio through his friend, Bruce Wang, who he used to work with.”

Kolton: Wong.

Andrew: Oh, sorry, Wong. But when we were trying to understand how you got the next batch of customers, all we got was that you iterated, improved the design, improved the user interface three times. I felt, like, our producer could have been a lot harder, like, to push more for more detail, especially since you did sales yourself. So now that I’ve got you on here, let’s more specific. What did you do to grow sales in the beginning?

Kolton: Yeah. I mean, I closed the first 5 or 10 deals. Expedia was on there, Warby Parker, a couple of other companies. You know, at Expedia, we’ve got a great champion, a great VP. Subbu, he’s very smart. He understands the space. He understands the tradeoffs. It was spending time with him. You know, what does this look like? How do you price it? You know, what’s fair? What’s appropriate?

Andrew: But then, let’s start fresh. You decide Expedia could be a good customer. How do you know that and then how do you connect with the right person there?

Kolton: Yeah. I think a lot of my first 10 or 20 customers, you know, they heard of us through a VC or they heard of us through the conference track. So I spent . . . Again, a lot of my early marketing was going to conferences and speaking and just teaching people. And that bore a lot of fruit. Expedia was, you know, my VCs had met this gentleman. They knew he was sharp. They’d thrown the idea by him. He was like, “Yeah, I’m interested in that. Let’s talk more.” But taking a interest in translating it into revenue is an art. For all the engineers out there, sales is harder than it looks. Marketing is harder than it looks. Give them credit.

And so, you know, then it’s, “Have you ever gone through procurement before?” Oh, snap. “What does your contract look like? Let’s redline terms of service. You know, how do you scale your pricing model? You know, how do we ensure we’re going to get a good deal as we grow?” And those are less, you know . . . I’m lucky because our product works and our product works well. I’ve talked to a lot of leaders that work for companies where the product maybe isn’t quite as fleshed out. We had the advantage of building it twice before we built it for everyone else. So I didn’t have to worry about, you know, does the product do what it says and can people use it? We had that down. So it was a lot more the business terms, the commercial terms, figuring out a lot of the details.

Andrew: If you saw my eyes flash everywhere, it’s because when you say, “I speak at conferences,” I don’t want to, like, take your word for it. I don’t know why. I like you. Why would you lie about that? But I still have to go and do research. And then I saw that you are speaking at an O’Reilly conference, that you are speaking at Chaos Conf. That’s your thing, right?

Kolton: Mm-hmm.

Andrew: Chaosconf.splashthat.com. QCon, San Francisco, also QCon, London. It’s you at all these places, right?

Kolton: Yep.

Andrew: And so when you speak at these events, people will then come to you afterwards or are you doing something to get them to connect with you or are you offering some reason for them to follow up? You’re smiling. There’s something there.

Kolton: Yeah. Well, the answer is, like, now we are. Back then, no. I would have some ad hoc conversations with people. I handed out a lot of stickers. And they would come to me or I would get a business card. And we’d follow up. You know, now we do these like sponsoring events in a booth and pre-event warming and post-event follow-up. It turns out, all of these are super important, but in the beginning, it was a little bit more fly-by-night.

Andrew: What’s pre-event warmup and post-event follow-up? I do a little bit of this, but I didn’t realize there was a name for it and a way to do it right.

Kolton: And I don’t know if those are the right names, but if you’re going to be at an event, you want to reach out to people that are local and let them know and give them an opportunity to attend. And if the people that you maybe have had touch points with in the past, they might be really interested in attending. So you want to give them that information. Post-event, you know, the game at events is, “Come scan my badge and I’m interested in you.” And so then what happens afterwards? Well, a lot of people get their T-shirt and ignore every follow-up email and that’s okay. We’ve got great t-shirts, but translating that into action. You know, did we find the right person in that company and do they have executive buy-in to go purchase a tool, are they feeling the pain, and are they ready, and are they mature enough?

And so part of that is how do we mature people? You know, if they’re interested but they’re not ready yet, then we have a lot of opportunity to provide education so that as they go through their career and as time passes, they’re getting little tidbits. Oh, you know what? I don’t need a full tool to go do this piece or how am I thinking about this process over here? And as they start to get those pieces in place, we’re building some trust. We’re showing them we can be valuable and we have useful things to teach them. And their organization’s getting better and they get to the point where, “Oh, you know what? I like that Gremlin team. You know, come teach me more. Come show me how to do this and [inaudible 00:50:13].”

Andrew: And does, “Teach me,” happen to be email or links or is it more than that?

Kolton: It’s event nurture is the term and that typically happens via email. People don’t know this. If you go to an event and you get your badge scanned, you’re going to get some emails from that company a day, three days, a week, two weeks, and it’s a nurture track. And if you engage with it, there’s one branch. And if you don’t engage with it, there’s another branch. And people can do that different ways. Some people hate them. I’m an inbox zero person. So I pretty rigorously delete those emails. But if it’s something you care about and it’s useful content and it’s valuable, and I think that’s the crux for marketing in the engineering world, [inaudible 00:50:55].

Andrew: Yeah. No, if you’re really interested, I get it, if this is a critical thing. Otherwise, I want to delete it just like you. I’m not an inbox zero. I’m an inbox ignorer, that same situation it feels like, but because of that, I underestimate the value of email as post-webinar. Not webinar, post-conference. I’m always thinking, I’m at an event, “Maybe we do drinks afterwards. Maybe we do dinner. Let’s do it right here.”

And I can see as I’m scanning all the different speaker notes for you, being a part of Netflix is huge for your credibility here. Being the guy who built FIT at Netflix, the failure injection service, is huge for you. And I can see how being associated with these companies, Amazon and Netflix, was a positive. Why don’t we do this? Let’s close it out with something that I should have asked you. You’re the engineer. You’re the guy who built the software. I haven’t asked you about the software much at all. The first version, how long did it take you? What did it look like? And then tell me a little bit about how you evolved it.

Kolton: Yeah. So the very first version was built at Amazon. It didn’t have a great API. I kind of lied earlier. The API needed to be improved, get cleaned up. It wasn’t really public-facing. I can show you a picture of the mascot, the graphic I created. It was an eight-bit version done in Paint. It took me less than a minute. Thank goodness I had a manager who [inaudible 00:52:12].

Andrew: On Gremlin or what you had at Amazon?

Kolton: What I had at Amazon. And so it was very humorous. It took about a year to build it the first go-around.

Andrew: When you left Netflix when it was time to do your own thing, Gremlin, you’ve got to start from scratch. Amazon’s not letting you take their stuff. How long did that take you and what was there because I kind of feel like having built something that was so good, it makes it harder to start from scratch?

Kolton: Yeah. So the quick side note, at Netflix I built the service and the API in three months and the user interface in a month a little bit later.

Andrew: Wow.

Kolton: So I did iterate better. At Gremlin, you get to make a lot of assumptions at big companies. Here’s what our infrastructure looks like. Here’s these toolings we have. Here’s how people authenticate. You’ve got to build that all from scratch. And Gremlin can be a potentially dangerous tool and so security and safety are very important. So it took us about a year. I think it was three to six months to build the client and the first version of the service and the API. It took us another three to six months to get the UI done. And then at about a year, I would call us, like . . . That’s, like, our alpha version.

And I sold that. We had some customers. We had some data, but it took another whole year before I felt like it was really enterprise-ready. And then, of course, you know, where are we at today? Another year’s gone by since then and now I feel like it’s great. You know, we’ve got designers and we’ve got dedicated security people and we’re going so much deeper. But a lot of the original code for the client and the service is still written by me and Matt and is still running.

Andrew: You heard me kind of wince as I was telling you about the . . . as I was watching you go to my website because oh, there’s one little change. We needed a redirect to bring the new home page up, but it wasn’t there. It’s hard when you build it and sell it sometimes, especially when you know how good it could be. I know how good this site will be, as soon as I tell Rebecca, “Hey, we forgot the redirect.” Was it hard for you to go out and talk to customers when you know how good it could be and it wasn’t there?

Kolton: It’s funny because one of my bosses at Amazon told me I had an unreasonably high bar, which I took as a compliment, but it can be hard. On one hand, the engineer in me is vocally self-critical, like, “Here’s what we don’t do well. Here’s what we should do well.” And I think that’s good in most conversations. It’s genuine. It’s honest. It’s credible. But, you know, so that’s one side.

You know, hearing customer feedback is another. Sometimes when people tell you your baby’s ugly or your baby should have three arms instead of two, you’re like, “I don’t know if I agree with you.” But that’s always a minute of reflection. And you’ve got to take that feedback and think through it honestly. You know, what do these customers know that I don’t? What’s different about their environment? What pain are they feeling? What problem are they trying to solve?

So that’s one where when you think you know the answer, and I’m not saying I do, but there’s a little bit of that in all of us. Then, you know, there’s some humbling experiences where you realize, “Oh, you know, we could do a lot better here.”

Andrew: I’m looking at the first version of your site. I feel, like, I don’t see the first product, but I do see the first marketing, the first explanation. You’re really good. I’ve been watching you. To be honest with you, I was a little afraid about having you on because you are an engineer in a leading role of a company whose product is not very, like, user-understandable. That’s not a good way to put it, but you’re so good at explaining it. Even from the beginning on your website, the headline . . . I was looking for a horrible headline to show how much you’ve grown. “Break things on purpose.” The subhead is, “Downtime is expensive and damages customer trust.” The call to action is super clear. Look, your lights went out again. That money. I get it. I get it. How did you get so good? Let’s close it out with that?

Kolton: I think it’s a life philosophy. You’re never good enough. You’re never done. There’s always more to learn. There’s always more to get better at. And it’s a hard lesson to tell yourself sometimes. You have to be humble. You have to recognize that you could do better. So I don’t know. I guess that’s one answer. Another answer is my mom wanted me to be a well-rounded individual. I play a musical instrument. You know, I was in marching band. I did this. I did that. You know, I know math and science. I know art. I know music. You know, I sing. Those are things where . . . You know, writing skills were very important to her and very important to my father, you know, how you present yourself, how you communicate to the world.

So I’ve just always tried to get better at whatever I’m working on. And, you know, that messaging, I spent a lot of time in my room. I could show you the chair that I sat on when I wrote most of Gremlin, when I thought through that messaging. I can remember sitting there and being like, “Well, this is catchy. Is it good? I don’t know. Let’s try it.” And really, it’s funny. It’s interviews like this. It’s those early customer discussions where you’re throwing out these tidbits and you’re looking them in the eye. Did they smile? Did they laugh? [inaudible 00:57:02]? Does it resonate? And from that, you start to distill what works. And then when it works, you run with it.

Andrew: All right. The website for anyone who wants to go check it out. What a good fricking domain. It’s gremlin.com. I love how I’m looking now at Internet Archive. There’s, like, gremlin eyes I guess, the green eyes that come out of nowhere. I love that.

My biggest takeaway from this is the lights going off as we’re talking. I think entrepreneurs would prefer coming into my office, especially when we’re in San Francisco, and we’re, like, two blocks down, not for you, but for some cases to go, “I’ll come into your office.” No, I want to see your space, your environment. I want to know that the lights are turning off. And in some cases, it bothers them so much, but I think that’s the best part. That’s where we get to see you, the guy who had his hand cut on a table when you worked at Amazon and now is refusing to find, like, a piece of tape to put on the goddamn lights to make sure that the lights stay on all the time. It maybe makes some entrepreneurs suffer to have us see that, but this is the part that I love the best, the part that is you when you’re not trying to be you and show us who you are.

All right. For anyone who wants to go check out Gremlin, go to their website. If you want your website hosted right, go to hostgator.com/mixergy. If you want to hire a phenomenal developer . . . I hired a finance person from them that I love and also a developer. So also a designer from them. Go check out toptal.com/mixergy. And yes, if you want to follow my run around the world, that was my goal from 10 years ago, I am going to fricking hit that goal. I’m so excited about it. I did 22 miles the other day. It nearly wiped me out. I was just done for the day, but I’m going to hit 26.2 miles all over the world, all seven continents, seven marathons in one year. Go check me out at runwithandrew.com. And I promise Rebecca will have that redirect looking nice for everybody.

Kolton, I’m looking forward to having a drink and find out what you did that was so wild.

Kolton: Hell, yeah. No, you’re a lot of fun to talk to, Andrew. I really appreciate this opportunity. We definitely . . . I’ll tell you the full unadulterated story.

Andrew: I would love it. Are you ever in San Francisco?

Kolton: All the time.

Andrew: All right. We should do . . . Do you play poker?

Kolton: I do play poker.

Andrew: But I do like $5, $10 poker. Are you too embarrassed to play $5 poker?

Kolton: No. I’m cheap. Remember, I’m frugal.

Andrew: Oh, right. Right. Right. All right. We’ll have to do it. And Adam, who’s listening in, thanks so much for setting this up. Bye, everyone.

Who should we feature on Mixergy? Let us know who you think would make a great interviewee.