decisions

How the Roomba Company Decided to Stop Making Bots for the Military

iRobot CEO Colin Angle with a Roomba. Photo: iRobot/iRobot

In 1990, fresh out of MIT’s Artificial Intelligence Lab, Colin Angle co-founded iRobot, a company best known for that friendly little robot vacuum the Roomba. In the early years, Roomba was unprofitable, and the majority of iRobot’s profits actually came in from military and defense contracts. That changed over time, and in 2016 iRobot divested itself its military and defense work to focus on robots for the home. Angle talked to Intelligencer about how he made the decision to divest from military projects, the difficulty in finding a new home for the defense division, and how his decision to focus completely on the Roomba means there’s now a robot vacuum that can empty itself.

iRobot was founded in 1990. The product that we’re most known for, the Roomba, came along 12 years after our founding. We knew if we could create a vacuuming robot it would be good, mostly because people kept asking for one. I’d introduce myself: “I’m Colin, I’m the CEO for iRobot.” And people would literally say, “Good to meet you, Colin. When are you gonna build that Rosie the Robot that will clean my floors?”

The challenge was, we didn’t know how to get there at any price point the average person could afford. It was a long journey that involved taking things from everywhere. The original algorithm used by Roomba to make sure it got every part of your room clean was taken from mine-hunting algorithms we had developed for the military. We built robot toys and patented them with Hasbro to learn how to build cheaper electromechanical devices.

At the same time, we had started working with the military. One early project we did was building an underwater walking robot that could detect mines in the surf zone. They were our crab robots, so we called them Sebastian and Ursula — though I’m not sure Ursula’s a crab.

In 2002, we both launched the Roomba and sent our robots to Afghanistan. The military robots were instantly profitable. They brought in something like $160 million in revenue and funded the company quite effectively. Roomba didn’t make any money for a few years, because we had so much to learn on how to manufacture, distribute, and support them. During this time, 100 percent of our profitability was coming from our business in defense with the government.

We went public in 2005, and the Roomba and military bots made for interesting bedfellows. We’d always get the question, “Why are you doing both?” And we’d say, “Well, because they’re both robots and the synergies between the kinds of robots we build for consumers and the robots that we build for defense are very real.”

But by 2014 we were starting to face real questions about how valid that was. Over the years, that supplemental military budget went away, so the revenue from the military went from $160 million down to $55 million. Competition in the defense business on the robot side was increasing. The cost of our robots was going to need to come down, the capabilities of the robots would need to continue to improve, and the next steps were getting expensive.

We were also seeing similar things with the Roomba, where competition was increasing. We wanted to make Roombas that could really navigate a room, and that required more dollars to be invested. Everything was getting more expensive and harder.

We even had issues around the branding of the company. What colors should the company use? Because they have to both be appealing to an average civilian and to someone in the military, with military aims in mind, and those are pretty different consumers.

The fact was we were increasingly two companies underneath the same roof, and we were continually making decisions where we had to prioritize one of these business units over the other.

But in 2014, we recognized that these were two very different businesses, and we were headed in a direction where we were going to have to choose. We couldn’t have had a great business if we invested in both home electronics and defense.

It’s important to point out how culturally important the defense business had been for our company. One of the things that really drove us was the impact that our robots have had in the world. We used defense tech to explore the Great Pyramid in Giza. We sent a PackBot to map radiation levels after the Fukushima disaster. With the Deepwater Horizon oil well disaster, we sent defense underwater robots to identify these giant subsurface pools of oil that were going to create a long-term ecological disaster. And, of course, during the Gulf War, we saved thousands of lives using the robots to defuse bombs and render IEDs safe. The idea that we could go out and have a real positive, chest-thumping impact on the world by deploying our technology, that was pretty central to why people felt good about being at iRobot.

On rational basis, it was obvious that one of these things had to go, and the growth prospects were on the Roomba rocket ship, while the defense business was contracting. And on the emotional side, oddly, how much I cared about the defense side of our business drove me to realize that keeping it was killing it.

We’d have these discussions, and then I would go home and really think about what I really wanted to happen. Which is a question I often ask myself when faced with a decision. Our defense side had done so much good in the world, so what I really wanted was for it to survive. So we hired a bank and we spent a year looking for the right place for it, and we just didn’t find it.

The next time around, there was that an ex-iRobot guy, Sean Bielat, who had worked on the defense side and then went to a private equity firm. He ended up convincing that private equity firm that iRobot should spin out our defense business and relaunch it. That made all kinds of sense. It meant that our defense side would be a startup again, and the new CEO would be a super well-respected ex-iRobot guy.

The devil was in the details, and it took about six or eight months to work everything out. Finally, Sean and Tom Frost — who was the internal leader of the defense business unit at the time — and I jointly announced the deal. That was very emotional, because that was the moment that there was really no going back. I think Tom and Sean realized what it meant to me. Their very sincere message to me was, “We will take good care of your baby.” They would make this new business not just about maximizing profit, but about maximizing the positive impact on the world.

The end game of the divestment coincided with this all-consuming proxy battle with an activist investor. He wanted iRobot to divest from defense, but he really didn’t want to hear from us that we were doing that. Some of the romance was stripped out of the experience by fighting this fight against a guy who wanted to replace our board and leadership, just so he could force the company to do something it was already about to do. That honestly was one of the more trying times. I’m trying to do the right thing by this company, and someone’s trying to come in and gut me.

It meant broadening my horizons around what is this thing that is iRobot. It’s a public company. It has shareholders that haven’t dedicated their entire lives to the business, nor do they have to, but they have every right to expect to be heard because they took money out of their pocket to be part owners of the company. That is what being a public company is about. I needed to be okay with that and figure out how to embrace that. I had to accept that this journey to build the company that I’ve been on for my entire adult life — there is no divine right for me to lead it.

I think I was rewarded for doing the right thing. Endeavor Robotics, which is what our defense business became, is thriving today. And investors in iRobot looked at the maturity of the leadership, its ability to make difficult decisions, and the overall good management during our proxy battle, and we ultimately won.

For us, being able to focus on the home market meant that in 2017, consumer growth quadrupled. We just launched the i7+. You can get a Roomba now and bring it to your home, and it will operate a year without your touching it, because it can empty itself — you don’t even have to pick it up. If you’re in the kitchen and make a mess, you can say, “Okay, Google, clean the kitchen.” And the robot actually now knows what the kitchen is because of the advancement in navigation technology. We’re still very far away from fulfilling the potential that robots have, but we just completely changed vacuuming, and that’s pretty cool.

If a company is like your kid, selling defense was like sending your kid off to college. Knowing that the people carrying the founding vision forward would do something great. That moment when Sean and Tom said, “We got this. We understand why it matters so much,” I knew letting go was the right thing to do.

If you look back at iRobot’s trajectory, and say, “Gee, what would you do differently if you had to do it over?” I think you could come up with a long list of things that’ve gotten us to a better place faster. At the same time, I say, “If I could travel back in time and do it over, I wouldn’t want to.” Because one of my most powerful assets was that I had complete ignorance about the size of the challenges that we were facing. I’m happy I was youthful and ignorant.

Why the Roomba Company Stopped Making Bots for the Military