Anthony built dnsimple on the side and didn't come on board full-time until after there were two other full-time employees. We discussed some of the advantages and disadvantages of running a complex infrastructure product, marketing a complicated business with comics, and fighting domain fraud through it all.
Garrett Dimon: Hello. Here today we have Anthony Eden, the founder of dnsimple, which is, as you would expect, DNS and registrar services. He is also the only person I’ve ever seen give a presentation knee-deep in the ocean.
Anthony, could you give a real quick 30-second rundown of how you got where you are in your career and what led to dnsimple? Just briefly without getting too much detail because we’ll cover the details here in a bit.
Anthony Eden: Sure. I’ve been a software developer by trade for many, many years, 20 years now. About six years ago, I just got fed up with using the domain registrar that I was using. I said, “OK. Enough of this.” I know I’ve been in this space for long enough.
I can build something that’ll work, so I built the first prototype and launched after about three months on the DNS side. It’s been growing ever since.
I started in 2010 and that was quite a while before I had the sense of how I was going to market it or do anything. I came at it from a typical developer’s perspective — I’m going to build a great product and the people are just going to show up at my door.
Garrett: Nice. How long did it take you to get full time on it, to leave behind everything else?
Anthony: It took about three years. I wasn’t even the first full-time employee. We hired two people in fact. Two other people were full time before I was even full time.
Garrett: You didn’t just jump into it, start cashing checks and running yachts?
Anthony: Oh, no, I wish it were that easy. [laughs] I started in 2010 and that was quite a while before I had the sense of how I was going to market it or do anything. I came at it from a typical developer’s perspective — I’m going to build a great product and the people are just going to show up at my door. [laughs]
What I learned very quickly is that it’s a long, slow slog, but it’s worth it if you have something that you believe in and that you continuously improve. That’s what I’ve been doing.
Garrett: One of the other things about dnsimple that really differentiates it is your challenge isn’t purely build a hosted piece of software and sell it. You’re dealing with infrastructure that’s critical. If it goes down it really wreaks havoc on your customers.
How has that affected of how you do things? Do you have any regrets creating an infrastructure business? Is there things you would do differently now that you’ve got a lot more of that experience under your belt?
…it’s challenging, but the reason I don’t regret it is because there’s also a moat. As you develop good infrastructure operational shops, and you have a team that can really operate things well, you actually get a strategic advantage, because not everybody can do that.
Anthony: Those are some great questions. Let me just start with the regret. Yes, it’s challenging, but the reason I don’t regret it is because there’s also a moat. As you develop good infrastructure operational shops, and you have a team that can really operate things well, you actually get a strategic advantage, because not everybody can do that.
There’s a small amount of people that can put together the right team, the right systems, and actually keeping it running. It can be exhausting, but at the same time, it gives you the freedom to do things that you couldn’t otherwise do.
Would I have done anything different? I think, we did a pretty good job of growing it from something which started off with a very basic infrastructure. Back in the very early days, it was actually run on multiple VPS providers.
Our nameservers were simple unicast nameservers that were running on a couple of different providers to provide some redundancy. It worked well for many, many years, and it gave us the chance to build up the rest of the things that we would use to provide value to our customers.
By the time we got to the point where we needed to expand our infrastructure to include things like an anycast network and stuff like that, which are quite a bit more expensive and difficult to implement, we actually had the capital to do it.
We again took an approach…It’s not like we went out and bought a whole bunch of servers, and our own cabinets, our own space, and datacenters. We went to managed service providers. I think, we were able to accomplish it because we stepped our way up slowly to the different levels of infrastructure.
Now we have an extremely solid bit of infrastructure, but it’s taken us six years to get here and a lot of negotiation, and working with other providers, and making things improve over time.
Garrett: It’s very different than running 1 of the 300 to-do list apps that are out there today.
Anthony: Yeah. There is the aspect of it that’s the SaaS side of it. There’s the running of the dnsimple application, and the API, and the other components. That has its own challenges as well. They’re just very different than the challenges from running a 40-node anycast nameserver system that’s spread across the world, then backing that behind DDoS protection services.
Getting everything to play well together is a big challenge. Yeah, there are some additional challenges. The good news is that, like many things, once you get them humming along, as long as you’re not adjusting things too drastically all the time, you can get to a pretty good steady state that makes it fairly easy to operate. Fairly easy, I say, in the big picture of things.
Garrett: You’ve also had to deal with fraud on a little different level. Whereas with Sifter, we dealt with a little bit and it was just some extra credit card fees, but in your case, when you’re registering domain names, there’s actually some significant challenges around that for you.
You don’t need to necessarily dive deep into the specifics of that, but is there any advice, any lessons you’ve learned from dealing with that, either from a technical standpoint, or a business standpoint, or just a sanity standpoint? What’s helped you manage all that?
Anthony: Sure, you’re absolutely right. There is a type of fraud that comes along from the domain industry that actually impacts domain registrars that that’s their primary business, fairly heavily, because essentially people can abuse domains.
They use them as a way to create phishing sites so that they can get more credit cards, so that they can then buy more domains to create more phishing sites. It’s the cycle that is so challenging to deal with. Our saving grace was that our business wasn’t just selling domain names.
The fact that we had, we built this as subscription service, and we offer more value beyond just the domain sales, means that the domain part of it was only a component of it, which gave us a little bit more of a safety net compared to if you’re just a pure domain registrar.
How we dealt with it was like everything thing else. When we started identifying those first bits of fraud, we would deal with it manually. We would try to track it down. After a little while, it became a point where I needed tools to help detect fraud.
Again, if you look back, our company was built in 2010. Even by 2012, 2013, when we started seeing this really take off, a lot of the fraud detection that was provided by credit card companies was subpar. They were protecting themselves from fraud, not really protecting the merchant from fraud.
Fortunately, the scenario has changed quite a bit today. In addition to new fraud tools from companies like Stripe, so directly from them, they’ve added new layers of fraud detections, and things like that, there are third party providers. The one we use is called Sift Science.
I highly recommend it. It’s actually done a really good job of helping us link together and identify fraudulent cases across multiple accounts and tie everything together. It’s still always a race. You’re basically competing to try to keep the bad guys from abusing the system too heavily.
At least, I now have some tools that help improve the situation and make it so that my job is not as difficult as it has been in the past. Frankly, though, if somebody is using the stolen credit card that’s completely valid and legitimate, you have to plan that into the business.
The card holders aren’t liable. The banks aren’t liable. It’s basically the merchants. That’s unfortunate because we have very little control over the security mechanisms that are on the cards themselves.
You have to plan for some level of chargebacks. If you don’t get them, then great, more profit. You have to factor that into the development of the business.
Garrett: Yeah, so it’s the cost of doing business. It’s not something that you can avoid. When I get that very first chargeback letter in the mail, it was Sifter, I was, “What the hell? Why is somebody doing this?” You reach out to them and 9 times out of 10, it’s a simple miscommunication.
In your case, this is people actively committing fraud and trying to…
Anthony: Yes. The other thing that’s frustrating about it, just one more quick thing. I would love to have…the thing that I miss is good, really healthy guidance from the banking industries and from the credit card industries about this. Unfortunately, merchants are the ones that are essentially left holding the bag in almost all these cases.
The card holders aren’t liable. The banks aren’t liable. It’s basically the merchants. That’s unfortunate because we have very little control over the security mechanisms that are on the cards themselves. My hope is that as banking continues to advance into the 21st century, that we’ll see more banks adding on value by adding more protections that are actually usable.
The other thing is you get these crappy systems like verified by Visa that are just a disaster. Instead, hopefully, we’ll get systems that actually work fairly well to stop fraud at the card level, which is where it should be stopped at in the first place.
Garrett: Yeah, for sure. I think that is definitely a feeling of helplessness as being somebody running the business when everybody else is protected and you’re the one that…There’s not a whole lot you can do. You can fight a chargeback, but in most cases, if it’s a $25 fee, it’s not cost effective for you to spend time fighting it, unless you have so many that you can afford a whole department to fight it.
Anthony: Exactly. Again, if it’s flat out fraud and somebody is using a stolen card, you’re still going to lose.
Garrett: Yeah, that’s true.
Garrett: On that note, and it may be fraud, maybe it’s different, what’s been the most painful day that you’ve dealt with running the business? Maybe it was a week, I don’t know.
I recall, 7:00 AM, still being at my computer after an entire night of trying to deal with this, and just ready to have a breakdown. That was really one of those moments, where I was like, “I don’t know if I could do this anymore.” I questioned my life decisions at that point.
Anthony: That’s actually pretty easy. The most painful day was December 1st, 2014. [laughs] I know the date because that was the day where we suffered a very significant denial-of-service attack against our service. It lasted for many, many hours throughout the night.
I recall, 7:00 AM, still being at my computer after an entire night of trying to deal with this, and just ready to have a breakdown. That was really one of those moments, where I was like, “I don’t know if I could do this anymore.” I questioned my life decisions at that point. [laughs]
The worst part is that I have a family. My kids were waking up to go to school. Seeing me still at the computer from the night before, they felt terrible about it. The impact was broader than just on me. It was affecting my family. It was affecting the employees. It was affecting our customers. That sucked. That was a really, really dark moment. It was a dark moment to the end of a very dark year.
Sometimes, everything has to fall apart. The good news is out of it, we managed to survive. I got a good piece of advice from somebody during that time. The piece of advice was, “When everything is going terrible, stop making decisions and just survive because you’re going to have to anyways.”
Once you’re at the other end of surviving, then you can come back and actually say, “OK, now what can we do to make good decisions to make sure this doesn’t happen in the future?”
Garrett: That’s probably a little more painful than most. I feel like so much of running a business involves fighting through that pain because I think everybody who’s ever run a business has had that day or week, where they’re like, “What the hell am I doing? Why am I doing this?”
It’s easy to give up and say, “You know what? This isn’t right. There’s got to be a better way.” Especially reading books by founders of huge companies, they’ve all got the stories. It’s a matter of how you fight through them, bounce back, that determines how successful it ends up being.
As an extension of that, is there a key turning point or a decision point that you feel like the business really either got over the hump and started humming along or just really changed how you viewed things and how things worked for you?
Anthony: It’s an ongoing thing. There have been pivotal moments. I can’t point to one specific pivotal moment. I would say that there have been some times where…
For example, when we went from being just two founders and one person, and one of the founders left to go do something else. That was one of those moments where it went from, this is just a little thing that I…one of those solopreneur-type things, even though there were two of us that found it.
It went from that to, “OK, this is a real business. I have to treat this as such,” which means developing systems for having employees and contractors and determining whether we’re successful or a failure, like, “Are we succeeding? Are we failing? What are the things that we’re measuring our success and failure by?” Things like that.
They came out of that change from that little company with just the founders and one other person, who also became a founder at that point, to a company that had contractors and employees, a larger amount of people depending on the survival of the company but who weren’t directly vesting…
They didn’t have direct vested interest in the success of the company, other than, “Hey, this is a company that provides us employment, and we enjoy doing things, and we feel like we’re making a difference.” That was a big change, that switch.
Garrett: The point in which you realized you had to start systematizing things and create a framework within which to grow the business, rather than just a business.
Anthony: Exactly, yeah. That’s a great way…The point of inflection was deciding it’s time to really start systematizing some of these things. From then on, that’s one of the things that I’ve been striving to do.
When I find these repeatable things in the business, I either systematize them myself or encourage others on the team to do them once. If you’re going to do them more than once, then write down what you’re doing and record that in something you could share with the rest of the team, so that in the future, others can help you out.
Otherwise, you’ll get in the situation where you have a team of bottlenecks, a team of everybody who has all their responsibilities and no way to share it. That’s unfortunate. That stops growth. That stops people from being happy because then they can’t have a vacation, for example, without feeling like they’re letting everybody down.
Garrett: Yup, absolutely. That’s one of the things that selling Sifter helped me realize. I knew it before, where it really beat into my head was selling, there’s value to systematizing everything for yourself, first and foremost, by writing it down and at least documenting your process, you’re helping yourself later.
There’s value in growing your team and transitioning people into new roles or hiring and all of that. It’s documented. You’re going to spend less time training them. You’re going to able to get them up to speed faster. They’re going to be productive and happier, faster.
Finally, when it does come down to it, if you do decide to sell the business, it’s going to make it a whole lot less painful to both sell the business and transition it to a new team. It’s one of those things where it’s just valuable, but it’s so hard. I think we all just get so heads down and caught up with the moment, we don’t take the time to invest in that stuff.
Garrett: Since you specialize in DNS and domain management, and since recently, I know having a secondary DNS setup has been a popular topic amongst people, after the recent Dyn incidents, is there any specific advice you would give to people who, maybe they run a blog, they set up DNS for their domain name, but they’ve never really run a larger application, where DNS can become more critical?
Is there any specific advice you would give to people like, “Make sure you’re doing this. Think about this. Research and learn about this”?
Anthony: Let’s see. I would definitely say that you have to think about failure cases. Often, when we first launch, we are focused on the success case, where it’s like, “OK, everything is going to go well. We’re only solving for that problem.” That’s fine. You should do that. You should do the smallest thing that you can do to begin achieving some success.
Then you have to also think about…especially once you’ve achieved some success, you found some product market fit. It’s time to start to think about continuity of business. It’s like, “How do I ensure now that these successes can keep going?”
I would say on the domain and DNS side, there’s a bunch of techniques, such as ensuring that you are the actual owner of the domain that you’re managing. In other words, don’t have somebody else register the domain.
Make sure the domain is registered on behalf of the business, that it’s clearly assigned to somebody, preferably to an email address list. Put controls over the top of it so that nobody can transfer the domain out without authorization. Use two-factor auth to ensure that it’s secure.
Make sure the domain is registered on behalf of the business, that it’s clearly assigned to somebody, preferably to an email address list. Put controls over the top of it so that nobody can transfer the domain out without authorization. Use two-factor auth to ensure that it’s secure.
Consider the domain like an asset that can be stolen from you as you would any asset that’s actually movable. A domain is effectively a movable bit of asset, so treat it like that. The same on the DNS side, you have to start to think, “Well, how much reliability do I really need? How much am I willing to pay for it?
Can I sacrifice some reliability in order to save some money and if I do that, what’s the impact later? Is it better if I pay a little bit of money and ensure that I have redundancy through two DNS providers for example? Am I willing to deal with extra complexity that comes around that as well?”
It’s not just always about saving money, sometimes it’s about system complexity and failure cases as they start to happen that you never expected because you don’t understand the complexity that comes out of it.
Garrett: It’s always ironic that adding extra layers to add resiliency increases more spots that can fail if they’re not setup correctly.
Anthony: Yeah, distributed systems are hard. [laughs] Really, really hard, and often, we underestimate the difficulties because the problems that occur aren’t easy to reason about all the time. We can reason about the execution of something that happens in a serial fashion.
It’s really hard for us to reason about lots of things that are happening in parallel and in timing that could be nanoseconds differences, makes all the difference in the world.
I would say the biggest thing that I see over and over again, unfortunately, when it comes to actual businesses, are the simple things of conflict between partners about who owns the domain name when there is issues with the business.
It’s a tough challenge. I would say the biggest thing that I see over and over again, unfortunately, when it comes to actual businesses, are the simple things of conflict between partners about who owns the domain name when there is issues with the business.
That alone accounts for some of the most challenging things that I’ve seen in businesses and other simple things like somebody registers the domain and then they just leave the company, but they have all the access to that. You have to treat it like anything else where you have multiple people have access but it’s controlled. You have to audit who has what.
You have to treat is as a true business asset. Many people, unfortunately, they don’t see their domains in that sense. They see it as something as like, “Oh, it’s just a domain name. I can get one for 10 bucks.”
Garrett: Yeah, it’s a lot more critical than that.
Anthony: That’s the problem. [laughs]
Garrett: I can’t remember where I saw it but somewhere it was in just matters of partnerships, people. It’s yes, dealing with all of that paperwork and detailing it out is difficult. Every day that it gets put off and something can go wrong, it will become exponentially more difficult to have those conversations.
Go ahead, fight through it, have them upfront, deal with it so that you’re securing the safety of the business in the future.
Anthony: If there is one thing I wish I could have had when I started dnsimple, it was a clear document. It was like here is a partnership operating agreement, [laughs] ready for you to fill in the blanks. This is a bullet-proof one that protects all parties involved and provides clear resolution when any member of the team leaves and provides clear evaluation of the organization and things like that.
Something that I could have had, I would have loved to have something like that. It didn’t exist unless you go to an expensive lawyer. The thing I learned over time, is that even expensive lawyers get it wrong. That’s part of the challenge. That’s one of the reasons why you often don’t find off-the-shelf documents like these that can be trusted because they just vary from case to case.
Man, there’s got to be a baseline document. It is like this is a normal operating agreement between partners that will protect all of you and that should not result in fights down the road, or if there will be fights, the boundaries of the fights will be clearly spelled out.
Garrett: It’s funny because its developers and product people, we’re all idealist and like, “Oh, let’s just build something and it’s going to be great.” In reality, so many of the things that kill businesses aren’t the lack of talking to customers. It’s these legal things that weren’t clearly defined ahead of time.
That’s the things that kill these businesses. Those aren’t the things that we’re thinking about and worrying about when we’re getting started because there’s too much enthusiasm, and excitement, and focus on the product. It’s a good reminder that people need to pay more attention to that when they’re getting started so that they can nip those problems in the bud.
Garrett: Switching gears a little bit, you all have got How DNS Works and you’ve done open-source work. You’ve got a handful of integrations with other tools that make life a lot easier.
From a marketing standpoint, would you say…Is there any particular strategy that you all found, has helped you reach customers and find new people that are interested in dnsimple or is it just a variety of things and you all are still figuring it out?
Anthony: We’re always figuring it out. There’s so many possible channels and ways to talk to customers and ways to position products. It’s often, if you position it for one set of customers, you might be leaving out another set. Early on, we used a lot of…I would go out to conferences and we would use talks.
We would use the things like that…We would talk to developers. Even today, still, our main way and we’ve really doubled down on lot of ways, our main audience that we speak to is software developers. These are the people that are running operations, that are building automation systems.
Because domains and DNS are such a core part of how the Internet works, we all have to touch it on a time to time. The vast majority of our customers are developers that want to, not just set it up, but in many cases, they want to do something with it that they could only do through automation, they can only do through the API and things like that.
Originally, it was going out to conferences. How DNS Works was an amazing success from the marketing side but completely unplanned. The designer on a team said, “I have this idea.” He sketched out the first two episodes and we were, “Oh, my God! This is amazing. You have to go through with it.”
He went all the way to the completion of the project. We obviously touched something in a lot of people because it has had a big impact in terms of the knowledge that people have of the company. Having said that, our bread and butter is still going out and talking to developers at conferences. It’s still building tools and good documentation around the API.
As we are getting closer to releasing our new version of API, we’ve been doing clients in different languages. Marketing through development is one of the things that we, as developers, find comfortable and we find that we can do it well, so we try to do that. We’ve been recently spending more and more time on content marketing.
The reason is because we saw evidence the most traffic site in all of the sites that we’ve launched it is, after How DNS Work had its big boom, but the most consistently trafficked site is our support site. This is the site that has all of our documentation that says, “Here’s how you do different things.” Once we realized that, we said, “Hey, maybe we ought to do some more work on the content.”
We started ramping that up a little bit more as well. Trying to bring valuable content that is both in the space of domains and DNS, but also things that start to stretch beyond that. People can understand that we actually have some experience running the business. They can trust that we know what we’re doing, and that we’re going to be here for a while.
Garrett: Getting started, was there a single source that your first customers came from? Was it more outreach? Was there a specific tactic or strategy that helped you reach those initial people or that it just happen over time?
Anthony: No. Our first customers came from my network and from the other founder’s network. When it launched, I went out on Twitter and I was talking to my network, which is the majority of my network has continues to be on Twitter. I would talk to them and I would say, “Hey, check this out,” and to get their feedback.
As they became customers, and at the time when we launched it, we were the early ones to say, “OK. DNS, specifically, at the low end but as a software, as a subscription, software as a service type thing.” There were a few other competitors, and they have been around for a long time.
We were starting this new…We were trying to jump into it late in the game. It was rebooting that idea a little bit. That worked for a while but that can only scale so much. We started working on getting out to conferences and talking to developers and growing it from there.
Really, my network was that first source of customers.
Garrett: OK, and so…
Anthony: People associate it…They said, “dnsimple is Anthony. Anthony is dnsimple.” That was how it was in the early days. [laughs]
Garrett: Definitely, when I remember when I first discovered it, I feel like it was pretty inextricable from you. Whereas, a lot of companies you run into it, and you might have an awareness, but I feel like your name definitely popped up a lot in those conversations, for sure.
I had another question. It’s not written down and I have lost it. We’ll skip that one. Looking back at the last six years now, I guess, little over six years?
Anthony: Yes, little over six years.
Garrett: What would you say has been the aspect of the business that has you excited to get out of bed and get to work in the morning or in some cases, not go to bed?
Anthony: [laughs] That’s a great question. After all these years, I still love the idea of innovating. I feel like in every space out there, in the world of computing and the Internet as a whole, we’ve only just started to see the level of possibility about what we can do.
In my little space of the Internet around the DNS world, I still see huge opportunity to improve the experience that people have, managing their domains, and to take away the fear and the nervousness around making changes and things like that. There’s so much more that we can do. That’s really what keeps me going.
I still am a hacker at heart. I love building things. Even though I’m running the business and I have all the business side of things, too, I still regularly hack on various bits of code because I enjoy it, because I love that creative aspect. That applies to the business as a whole. I still see us as a creative, innovative endeavor just in our particular space of domains and DNS.
Garrett: As a founder, how do you balance that desire for creating new things and exploring and having fun with the need to help the business stay reliable, sustainable, and afloat, and handle those boring things that end up consuming so much of your time?
Anthony: Again, the thing that I try to do is with processes.
I try to get us to develop us as a team, so the entire dnsimple team, to think about what they’re doing and convert it into a process that I can understand and that the rest of team can understand, because then, I have less need to actually think about the operational aspects of the company. I can go in and focus on the creative side of the company.
Now, there are other aspects. It’s tough to work with other people many times. Making sure that people have a healthy and happy environment where they communicate well with the rest of the team, and where they work well together, and that they’re all productive, is a challenge.
We’re all different people and sometimes feelings get hurt, sometimes things get said or don’t get said that need to be said. I still see and I dedicate a particular amount of time in each my week to either talk directly one-on-one.
We do one-on-ones every two weeks with everybody on the team or I will set aside time to talk with the team as a whole. I will say, “OK. We need to deal with something,” so, I’ll schedule a specific amount of time to talk to one or two people. I consider that a part of my job.
My job is to make sure that I help the team work well together. Sometimes, I’ll just hack at seven in the morning when I wake up. Before anybody else is really on, before I have to do the other stuff, I’ll throw an hour of code out there just to see what happens or I’ll do some research on something that I think is interesting. To me, that’s satisfying.
I like my job so much that I don’t consider it a job. I just consider it a part of my life. It just blends right in. Flipping open the laptop and hacking for a couple hours if I have some free time, is something I’ll do whenever I have it just because I enjoy doing it. I don’t need to sacrifice the operations of the business.
I will always try to make sure that goes first, but I can still do that and find some time here and there to hack on new ideas or to propose things to the rest of the team, or what have you.
Garrett: Yeah, right on. This brings us to the end. Is there any simple parting words of advice you would give to anybody else that wants to start a software business, hosted or otherwise, based on what you’ve learned?
Start small, release early, release something quickly, learn how to sell just the small thing you have, and iterate over the development of the business as a whole, of the development of the product, and the development of your marketing strategy.
Anthony: Start small, release early, release something quickly, learn how to sell just the small thing you have, and iterate over the development of the business as a whole, of the development of the product, and the development of your marketing strategy.
Just like you would on anything else that you’re writing as a bit of software, iterate on the business as a whole. That’s probably your best way to reach success, I would say.
Garrett: Right on. I would completely agree with that. I think too many people start out too ambitious, bite off more than they can chew, and then end up collapsing under the weight of just how big those projects can be once you start getting into it.
Fantastic. Thanks so much for doing this, I really appreciate it. There’s some great stuff in here.
Anthony: It’s my pleasure.