The founders of Mailgun and Gravitational on the ups and downs of selling your startup, and going back to entrepreneurship (and YC) again.
For weekly recaps of The Macro, sign up here.
In 2008, Taylor Wakefield started a startup, and brought Ev Kontsevoy on as CTO. Hindered by the financial crisis that year, that company failed -- but out of it came Mailgun (YC W11). Sasha Klizhentas joined Mailgun during its early stages as a back-end engineer and scaled it from 100 users to 100,000 users. After being acquired by Rackspace, the three are now building a new company, Gravitational (S15).
The Macro : Two of you went through YC in 2011. What was that like compared to your second time around?
Taylor : Whether it’s your first time or your fifth time, I think you get a lot of value out of the program. By the nature of what it is, you probably get the most value the first time but the second time I was shocked by how much the network of investors, partners and participants had grown. Going through YC again expanded our reach a lot. Plus, there’s always a lot of friendly competition around you during YC and that makes you work harder, and achieve things faster.
Ev : In 2011, there were 42 companies in our batch and somehow we managed to meet everyone and become friends. Today, the network is much bigger, which actually made it a little more challenging to meet everyone. There are also more YC partners, and their experiences have gotten much broader. Now, there’s a much more targeted conversation with a partner around their certain expertise.
The Macro : What made you decide to sell Mailgun to Rackspace?
Ev : I think that’s a very personal question. You will probably get three different answers from each of us. In my mind, there were two components at play. The first was a good financial outcome for my investors, employees, and ourselves. And the second one was the moment when I could clearly see what I would be doing at the Rackspace. I had a very frank conversation with Rackspace’s leadership about wanting to build bigger products, things that startups aren’t normally funded for. I wanted to get into hardware and learn more about enterprise sales and supply chains. It provided that opportunity so that was why it was pretty clear it was going to be a good acquisition.
Taylor : At the time, we were at a crossroads: We could have turned into a lifestyle business and grown gradually over time; we could’ve tried to raise a Series A; or we could have gone down our current path. I think there was a personal financial consideration, as well as whether Rackspace was going to support the brand and product. But we also had some limitations. We were running out of IP addresses. We were trying to get Green Cards for our engineers in Russia. So there were just a lot of practical considerations. Rackspace was going to make it a lot easier.
Sasha : On top of that, we also thought a lot about our team. At Rackspace, there were technical career tracks that would allow our engineers to do things we couldn’t offer as a smaller startup. Knowing that there were opportunities for them was really comforting when we talked about the acquisition. Every one of our engineers stayed at Rackspace for 3 years or longer. Everyone was pretty happy and got to do things that they didn’t do before. A lot of companies actually have different tracks for engineers, so knowing how your team can integrate into that will make the process a lot easier.
The Macro : What would you tell founders who are thinking about acquisitions?
Ev : You have to be very honest with yourself as a founder. Are you building to sell, or are you building to dominate? Because if you fail at becoming a dominating monopoly, then an acquisition will be more about accepting defeat and minimizing damages. Looking back, it's clear that from the beginning Mailgun was built to be acquired. We did not see how it could ever become a billion dollar business on its own. So you have to be at peace with the fact that your product will eventually be owned by another company. We just had to find a good candidate that shared a lot of our values and we did. If you don’t, all the other stuff -- like term sheets and negotiations -- become a nightmare. One mistake that we made was not getting our investors involved early on. We would have saved ourselves a lot of time and money if we had brought them into the conversation.
Taylor : Acquisitions can take a long time, so be prepared to spend the next 6 to 12 months on it. It’s pretty distracting when you’re running a business. But you have to maintain growth.
And you have to be aware of whether or not you have leverage, because if you don’t, you will get screwed over during negotiations. With most negotiations, there’s going to be a point where you have to be willing to walk away in order to get what you want. At the time, Mailgun was profitable so we could’ve walked away and still not be out of business. It also helped that we almost doubled our revenue between when we first started talking to Rackspace and when we closed the deal. That growth gave us continuous leverage during negotiations.
Sasha : From an engineering perspective, one thing to keep in mind is technical diligence. At Gravitational, we started a spreadsheet that keeps track of every license we use in our code. You have to research the licenses that you use because your customers and any acquiring company will want to know that your IP is clean. If you don’t keep track early on, it’ll be very difficult later on to retroactively find or even modify code. It’s easier to address every pull request than to dig through an unmanageable license soup. It becomes a headache when you’ve accidentally violated a license and need to rewrite a core piece of your product. We’ve luckily never had this issue, but it does introduce a lot of stress and wasted time.
The Macro : How did your time at Rackspace transition into starting Gravitational?
Ev : Our time at Rackspace gave us a lot of access to some of the pain points of enterprise clients. We saw that engineers generally engage in two different types of activities: software creation and software maintenance. While the market for tools for the former has been oversaturated, the latter has actually been somewhat neglected. Companies struggle as they scale because the problem shifts from software creation to software maintenance. In a team of 50 engineers, 40 of them will be scaling and maintaining the software. That’s the reason why companies struggle to scale to multiple sites. It’s not that deploying code is hard; it’s actually very easy. The hard part is scaling your operations and this is why they get locked in into one AWS region or one data center and cannot sell on-premise easily.
Companies started offering us money to solve that problem for them, and so we realized there needs to be a company that does this professionally. We started Gravitational to help companies run their software on thousands of locations.
This tooling also allows companies to make a cleaner separation between development and operations. The role of DevOps has translated into making developers do everything but we think operations should be a unique job. Developers write code and operators maintain the applications. By separating the roles and creating tooling specifically for operations, we allow companies to scale better, like Google. It’s not a coincidence that we leverage Kubernetes, which was built by Google. An added benefit of being able to run software on thousands of locations means you can run it on your customers’ infrastructure. This is usually a requirement for large enterprise companies. So in that regard, Gravitational helps bridge the gap between software vendors and enterprise customers.
The Macro : Were there challenges the second time around that you didn’t realize would be challenging?
Ev : There’s a big difference between our past startup and what we do now. Our customers now are much larger and so we’re solving much larger problems, making our past experience less applicable. There’s quite the learning curve, but it’s been a lot of fun because there’s a human touch to it. We’re no longer just sitting at a desk and building a product. We spend more time with our customers and see how our product is improving their lives.
Taylor : Because we are dealing with enterprises, the sales cycle has become much longer. There are times when you think a deal will slip and you have to rescue it by negotiating and compromising. You don’t really experience that when all you have is a signup form on your website. It’s been an interesting learning experience.
We also have three co-founders now, so communication needed to be better. Every time you add an extra node, communication gets a little more difficult. It’s important to be cognizant of what each person’s role is and how that all works together.
Sasha : You take certain things for granted when you’re at a bigger company with an established brand. When you launch a product, there’s a line of journalists ready to interview you and talk to you about what you’ve been doing for the past 6 months. It was kind of an unexpected shock going back into the startup mode. Even though we’re building much more impressive technology, we actually have to fight for media spotlight.
The Macro : What advice would you give to the incoming batch?
Ev : This is a very common piece of advice but not a lot of people take it seriously: Be easy on yourself. At the end of the day, the reason why people start companies is to make other people feel better. People don’t just buy software, they buy software that makes them feel better. So if you’re in the business of making people feel good, why don’t you make yourself feel good too?
Sometimes founders neglect that by pushing themselves too hard and often it’s just unnecessary. Treat yourself and your employees as you would your own customers.
Taylor : I’d say to take your time at YC as an opportunity to meet a lot of interesting and smart people. It’s a pretty amazing opportunity. The one thing you won’t be able to do after you leave is meet these people face-to-face. That’s something we didn’t execute as well on during the last batch as we should have.