Welcome! This week we’re talking with Colin Angle, the CEO of iRobot, makers of the hugely popular home vacuum cleaner Roomba, as well as a number of other autonomous devices. This interview is a big one for me, because I’m obsessed with my Roomba robot! In iRobot’s early days, Colin and his team designed the behavior-controlled rovers for the National Aeronautical and Space Administration that led to the Sojourner exploring Mars in 1997. He holds a B.S. in Electrical Engineering and an M.S. in Computer Science, both from MIT.
Welcome, Colin! I grew up reading a lot of science fiction novels, so I have to ask — was the name iRobot inspired by the Isaac Asimov novel?
Well, according to all the books and movies I grew up with, today we should have robots everywhere, right? And they should also be way smarter than they are today! The original idea for our company was to build robots that had human-level intelligence, but we knew that real AI was going to take a while, so we decided on a hack. At the time, the Internet was gaining momentum, so the idea was to take advantage of all this new connectivity and build robot avatars for people, who would be providing the intelligence until the AI caught up. So the “i” actually stands for Internet. We were also fans of Asimov, of course.
How many robots have you developed over the life of the company?
Well over one hundred. We’ve been building all kinds of robots for over thirty years now. We entered and exited more than 20 different robot businesses before we got to Roomba, which is essentially a mash-up of a bomb disposal robot, a toy robot, and an industrial cleaning robot. We said, “What if we put all these technologies together to make an affordable robot for your home? Wouldn’t that be cool?” Today we are one hundred percent focused on the home and the consumer. To date, we’ve sold over 30 million robots worldwide.
Now we certainly still aspire to all sorts of diversity in our product line-up, but we’ve also realized that we need new ways of selling. We want robots to be everywhere (that’s what we were promised), and in order to make that happen we had to continue evolving the business model. We used to sell everything through traditional retail channels, but we also needed a more direct, affordable solution. So subscriptions and the direct-to-consumer model are a huge unlock for iRobot to continue repopulating the world with the cool robots that we were promised as kids. The idea is to make smart homes that are actually smart.
Which brings me to my next question: Why is my smart home so dumb?
The way the smart home model works today, we slap wi-fi chips on everything and wind up with 72 different apps on our phone. We might use a particular interface once or twice, and then promptly forget how to use it again. Here I would take it back to this idea of re-imagining robots. We’re so used to thinking of them as these autonomous machines, from Robby the Robot to R2-D2. Well, what if your entire home was a robot? What if all these cameras and sensors were simply the eyes and ears of your home? Maybe the doors and window shades are the arms and legs. And maybe the Roomba is like a white blood cell that’s shuttling around trying to keep things in order.
Who says that a robot has to be a single device? By rethinking what a robot is, we can imagine a scenario where you live inside a robot home that can understand and collaborate. It can speak your language. That way you don’t have to remember all those dozens of apps with their own unique commands and operating procedures.
Sort of like Jarvis from Iron Man.
Exactly. When Tony Stark accidentally sets himself on fire, Jarvis is there to save the day by activating a nearby device. And again, this idea fits really well with a service model, because you probably wouldn’t want to buy all this stuff at once.
Imagining your home as a service gives both the consumer and the vendor that much more confidence to create a great holistic solution, as opposed to buying all this stuff piecemeal and running crazy science experiments getting it all to work together. You just can’t deliver solutions at scale with the current retail model.
But the alternative is really simple and great. You pay a certain amount a month, and then a box of stuff shows up, and it works with everything seamlessly, and every time you plug something in your home gets that much smarter. You don’t have to program anything. It just works.
When I buy a Roomba, I’m obviously not just buying one device. I’m buying the collective intelligence of all the millions of Roombas out there. How are you using that intelligence?
I agree, you are buying the collective intelligence. In fact, iRobot has the largest deployed network of machine learning devices in the world, other than Tesla. And every quarter, we do a major upgrade in the algorithms and what the robots can do. They get better all the time.
Let’s say your Roomba gets stuck somewhere. Well, is there anything we can see immediately prior to getting stuck that we might be able to recognize as a red flag? Or if we identify a particularly problematic place in the house, well, why don’t we just avoid that spot? Those are the sorts of issues and problems we’re solving all the time.
We work with a lot of companies that are launching connected devices. What’s the one key piece of advice you would give them? What’s the biggest mistake to avoid?
In the land of all of the mistakes that I’ve made in my quest to create practical robots, one of the bigger ones was this idea that the perfect Roomba was the Roomba you never saw and you never touched. You just came home every day to a perfectly clean home. Boy, was I wrong.
And it only took one customer to explain it to me. And he said, “You know, Colin, imagine you had a cleaning person that you couldn’t talk to, and you had to kind of trust that they would do the right thing. Would you trust them, and would that be a good experience?”
And my response was, “Oh, yeah, no. That doesn’t sound very compelling.” We need to move beyond autonomy towards collaborative intelligence. The big idea behind Roomba is to create a robot that will feel like a partner; that will be a good listener and speak your language, the language of the home. It will know what the den is and what the bedroom is and what the kitchen is, or whatever you call your different nooks and crannies of your home.
So you really need to be thinking about your device in context. It’s actually the beginning of a new kind of relationship with your customer. Did you look at any other devices or subscription business models for inspiration?
That’s where your book was very helpful. It gave us some concrete examples to work from. At the end of the day, we want to sell man-machine partnerships based on differentiated experiences. So looking at an example like Fender was really interesting, because at the end of their day, they want to sell musicianship, not just guitars.
I also gave it to my leadership team so that it would give us a new financial language. It gave us the math, as well as examples of other companies that have successfully made this shift, so that now we are able to start measuring the things that are going to earn us the right to change our behavior from transaction-based to relationship-based.
It’s the difference between a CFO going from “Yeah, not in my lifetime,” to “Okay, you want to do this, Colin, here’s what you need to show.” Because when you’re thinking about increasing customer lifetime value, and you have confidence that your churn is low enough that you can amortize the cost of customer acquisition over time, and you can make credible statements around multi-year paybacks on customer acquisition, then it changes everything.
It’s a new kind of financial language and a new kind of collaboration. Sort of like this new kind of partnership with robots that you’re creating.
Thanks so much, Colin!
Any time, Tien.