The iRobot Q&A: Colin Angle on the Next Robotics Inflection Point
It is a time of change at iRobot. Ten years after the landmark release of Roomba, the robotic vacuum cleaner, the Bedford, MA-based company (NASDAQ: IRBT) could be considered the most successful robotics firm in history, even as it is trying to reinvent itself with new product lines.
This reinvention comes amid big uncertainties in the U.S. defense budget—a longtime staple of iRobot’s revenues, which will top $450 million this year—as well as huge new opportunities in areas such as healthcare.
Indeed, the company’s newest class of robot—unveiled last month in partnership with InTouch Health—is built specifically for the healthcare industry. The two parts of its name, RP-VITA, are meant to convey what it does. RP, for “Remote Presence,” means it will help people interact with others when they can’t be there physically. VITA, for “Virtual + Independent Telemedicine Assistant,” means it will be used in hospitals and clinics to help doctors treat patients, and that it can navigate around by itself. (OK, the name might be a little clunky, but hey, C-3PO didn’t roll off the tongue right away either.)
On a recent visit to iRobot headquarters, I sat down with CEO and co-founder Colin Angle for a wide-ranging interview. We talked about the company’s new robot, what inspired it, and its significance to the field of healthcare and to iRobot’s future. Angle also reflected on the ups and downs of building a company in the robotics industry, and how the field has evolved since iRobot started back in 1990 (originally called IS Robotics). He expounded on lessons he has learned, some current issues and threats in the industry, and what it all could mean for the future.
In the end it’s refreshing to see that Angle, a tech celebrity and an iconic figure in Boston’s innovation scene, hasn’t lost his boyish love of robots. (None of these guys ever do.)
Here’s a lightly edited transcript of our chat:
Xconomy: When people think of iRobot, they think of Roombas in the home and PackBots in the military. Where did the healthcare robot come from?
Colin Angle: It’s been a long-term interest for me. When we did our first telepresence robot back in 1999-2000, we felt like if the technology could deliver a sufficiently compelling experience, then this [healthcare] was going to be an important application. The problem was, Wi-Fi was not ubiquitous, the bandwidth was not there, the systems that were built were more, I would call them “minutes of fun,” than useful tools for collaboration.
X: So the technology has really come along in the past decade.
CA: One of the things that came out of the defense research and development is the ability to do practical high-speed navigation and mapping. “Aware” software is our robot intelligence system that drives our robots. Integrating that operating system, with Visual SLAM [Simultaneous Localization and Mapping], which uses cameras as the primary sensor. Our advances in that technology started to create the opportunity to do more than tele-operations. Soldiers didn’t want to be staring at a screen getting their robot into position when somebody might be targeting them. Finally, we have practical robots that understand enough about their environment to know where they are.
That led to a very exciting technology demonstration called Ava. The architectural concept was, from the neck down, we were going to have a robot that could do navigation mapping, obstacle avoidance, and from the neck up we were going to put a tablet [computer] to tap into the mobile computing industry. Getting voice and video over IP, sexy touch screen interfaces, powerful mass-market trained development tools, speaker independent voice recognition, facial recognition, all of the stuff the mobile industry is doing, and get that “for free.” So we launched that a couple years ago.
Now we’ve just launched this [RP-VITA robot (see left)], which is the first practical application of the Ava technology, though far from the last. It’s done in partnership with InTouch Health. The mission of this robot is, we want to start at the high end. We have great confidence in our system—we wanted to demonstrate that not only could we navigate and provide a great remote presence experience, we could do it in a hospital. The good thing about a hospital is you’re leveraging a doctor’s time, which is a very valuable commodity. But the technical challenge is that you’re operating in medical environments, which are high-consequence if things go wrong. We’ve been able to create this system that will navigate in hospitals and allow this to be an avatar for a doctor and actually do medical diagnosis remotely.
X: So what’s the deeper idea here?
CA: One of our goals is to create a system—this might sound radical—that’s better than being there. What would it take for a remote doctor using this system to compensate for his lack of physically being there with information and technology? If I’m looking at you on this system, I can pull up your medical records, get real-time displays of all the systems you’re hooked up with, pull up data associated with you. I can also know who your treatment team is, and just by pushing some buttons, pull them into this robot so we can all look at you and have a conference call and discuss what should be done.
This integration of information with this high-quality system which can get itself around the hospital on a schedule—visit this doctor, this patient, that patient, monitor the patients—is really a holistic patient care system to allow more expertise with the right information to be brought to bear at the right time against an array of patients that need that care. It’s not whoever’s in the building, it’s whoever in the world is the most appropriate person or team to treat you when you need it. That’s the big idea as far as how this particular system will create value to the team. We’re really excited about this. The next concept is a more generic remote presence system that will allow video conferencing to be used more aggressively in more settings as well. So we’re just getting started, obviously, in what we think is going to be an important new venture.
X: Can you talk more about the broader impact? How will this specifically affect the quality and cost of patient care?
CA: First, what will happen in hospitals—it’s already happening—procedure by procedure, application by application, robots will be used more and more. Right now robots have some real inroads in stroke, where there are two main types of stroke. One type is treatable by t-PA [tissue plasminogen activator], but if you give it to the wrong kind of stroke, it’s detrimental—so you need a specialist to make the diagnosis, and they’re present at some number of city hospitals, but not in your local hospital. That diagnosis needs to be made within three hours of onset of stroke in order to maximally have positive effects. A robot would allow a specialist to see a patient in a rural hospital, make the diagnosis, and be a win-win for everyone. The local hospital can treat the patient, gets the admit; 30 percent of the admits would ultimately need to be transferred into the big hospital, so the big hospital gets the referral business. As you play this forward, you can improve the quality of care for the patient by allowing the right expertise with the right tools at the right time at the patient, through ubiquitous access to telemedicine tools. So that’s a huge win in patient care.
And then as you go forward, there’s an even bigger idea. How do we treat people to an increasingly significant degree in their own homes? Because one of the big problems around the cost of care is the cost of the hospital. Ignoring completely all of the capital investment that people have in their own homes, the whole care network with your siblings and parents and children that is available. This type of technology ultimately will allow doctors and nurses to make house calls in a cost-effective fashion. Which will mean that you go to the hospital only if there’s something that only can be treated in the hospital. This will have massive impact on the cost of care and quality of care that patients enjoy. So I think this is a critical bit of technology. Especially as our aging demographic is going to be placing demands on our healthcare system. We need this stuff to exist. Hopefully we’re not too ahead of the curve. I don’t think we are, I think the curve is there right now if the products and technology and price points all line up. So, together with our partners, we’re on a path to create the systems required to deliver that future.
X: Of course, watching the robot move around and interact with its environment is still the most impressive thing, at first glance…
CA: If there was an appropriate price to live in your home, this would be the robot that met you at the door, talked to you, and would have speaker-independent voice recognition. You could convey your will to it, and its mission would be to activate Roomba, tell Roomba where to go, activate Scooba [floor-washing robot], tell Scooba where to go, and so forth. But you can see that it’s fast. You can step in front of it and do whatever you want to mess with it, and it will dynamically respond to what’s going on.
X: Let’s shift gears to broader issues. What’s the biggest lesson you’ve learned over the past 22 years?
CA: So much about the challenges that this industry faces is a recognition that it needs to be business-driven and not technology-driven. There is more technology than there are good business plans. So much of the work that we need to do as an industry is finding the places where the technology can actually create sufficient value to drive viable businesses. The Roomba was launched in 2002. From a technological perspective, it could have been launched earlier. The challenge was did anyone know and live and work through the business that allowed this system to be sufficiently easy to use at price points that retail distribution could support?
That challenge is still with us. This is a complete disaster and waste of money if you’re just doing it to create a cool demonstration. Whatever you do, you want to look at the business underpinnings. That doesn’t mean you only do what people say is a good business. With our healthcare initiatives, there are trends that we’re working towards. But we at least know we are attacking real problems, and if we can solve them we’re creating economic value.
If I had a firmer grasp of the business side when I started in 1990, we would have seen more progress more rapidly. We managed to survive long enough to develop an appreciation and a competence, and our mission—build cool stuff, deliver great product, make money, have fun, change the world—was unique back then because we explicitly talked about making money. Everything we did had that dimension to it. Ultimately that mission, and that statement of the mission, gave us the motivation and insight to actually do it and allowed us to make more progress.
X: So have things changed dramatically in the robotics industry, from a business standpoint?
CA: If you look at the robotics landscape today, it’s much richer and valuable robot businesses because more and more people are approaching it from the perspective of [here’s a] problem—solve it with robot technology—as opposed to, “Woo-hoo, look at my walking robot.” That’s great, that’s very exciting for me as a lifelong supporter of this industry. Sometimes I feel like I’m talking to a wall that doesn’t understand, but it is very gratifying to see the increase in quality of businesses out there today.
X: What were the most important inflection points in the company’s growth?
CA: The biggest inflection point in iRobot history was 2002. iRobot was basically flat. In 2002 we did $13 million of revenue. The year before we had done $10 or $11 million, and the year before that we’d done $10 or $11 million. We were doing this breadth-first exploration of value-creating opportunities for the robot industry. 2002 started with our first layoff in the company’s history, and ended with putting a robot into the Great Pyramid in Giza, launching Roomba, and sending PackBot to Afghanistan. We went from $13 million to $42 million in revenue in 2003, and then to $92 million. That was an inflection point.
The magic moment of that whole experience was a phone call with Brookstone. There was a buyer at Brookstone—she called me up and said, “Just want to check in with you, Colin.” Oh, how are we doing? Are we doing OK? “Yeah, Colin, we’re doing OK.” Is the sell-through good, do you have enough inventory? “Actually, that’s why I’m calling.” OK, well how many more Roombas do you need? And she said, “How many can you build?” That was the moment when we knew that we had something. We had set off to sell 15,000 Roombas. And in the third quarter of 2002, we sold 70,000. It forever changed the business.
X: Now, with your business diversifying, could you be headed for another inflection point? What will that take?
CA: The mainstreaming of Roomba is very exciting. That is something that hasn’t been seen before. You get market data back for the small domestic appliance industry. The industry has grown 3 percent this year—this is toothbrushes, coffee makers, irons, and so forth, upright vacuum cleaners. Why? Because of espresso machines, robot vacuuming, and single serve coffee makers. The first and third are a pre-existing large industry. Robot vacuuming is still relatively small, but it’s growing so rapidly that it’s moving and growing the whole industry. So that’s big.
On the defense side, we’re being negatively impacted by the uncertainty around sequestration, the election, and budgets. It’s not an inflection point for the defense [business]. In the remote presence industry, I think that this [healthcare] robot is important. Is it the beginning of a revenue hockey stick? It’s going to take a little while. We’ll do between $465 and $485 million this year. It’s going to take a little bit in order to materially move that number with a new-to-the-world product. But we think it will be material next year and we think it will be very, very important over the next few years. So from a financial materiality perspective, a little more patience before you see that curve, but from a first practical navigating remote presence system, I think we’re going to trace back to this system right here as being incredibly important in all of the new economic opportunities that navigating, mapping robots are going to create for the world.
Roomba is “clean the floor completely, you’re good.” This is “go to the room swiftly and safely, and perform value.” It’s a different class of economic opportunity. You could spend five minutes and list out more applications for this technology than any one company could develop in its lifetime. So that’s cool. We need to focus systematically, and grow from one to the next.
X: Let’s talk about the competitive landscape, whether it’s Google or other Silicon Valley companies. Who’s going to own the operating system for robots?
CA: Right now, iRobot has the “Aware 2” system, architected from the ground up to be reliable, safe, secure, very efficient, and we’re an open interface as opposed to open source. That has allowed us to create the navigation demo—it’s very fast. Our overall strategy is one where we’re going to embrace mobile from the neck up, and create a safe system from the neck down that can handle the mobility and areas where crashes in unsafe fashions can lead to very bad things happening. We have a commercial-viable strategy.
I think the Willow [Garage] folks, it’s more of an academic-friendly, open-source environment, which plays well to the needs of the academic community and has a long-term role in the ecosystem. Can it ever be sufficiently reliable and capable for commercial systems? Maybe. Certainly there are some startups using Willow. My firm belief is that more companies is good for the industry. My concern around an open software system like Willow is that we need to make sure that the [intellectual property] that enables robots to perform whatever value-creating business that the companies are out to create, that that IP is ownable by the company that is developing it.
Because if you think about the software industry, there was enough IP ownable to allow application development on standards to create value for the founders. In the robot space, the more value that ends up in the physical hardware, the less of a defendable competitive position you have. Because ultimately, unless you’ve done something really wonderfully creative with your physical system, the hardware is more vulnerable to being commodity than the software. So my anxiety over Willow is that we commoditize the value of the software, forcing robot companies to compete on hardware, which gives the advantage to big giant [consumer electronics] firms over in Asia, which facilitates the exporting of the robot industry overseas. That is a macro anxiety that I have. But the positive side of Willow is that it’s a forum for academic innovation. We look at the algorithms, and what’s being done at Willow, and have taken elements that we think are valuable and pull them into our arena and use them.
X: Looking back on all your years at iRobot, what’s the biggest surprise?
CA: That it would take so long. This is a hard industry. It takes a lot longer to travel from idea to product than I would have ever guessed. Another is that the difference between a demonstration and a product is much larger in the robot space. The difference between demo and product in the software space can be weeks—or it’s the same, you just have to deal with it crashing.
When we started the business, we thought robots were cool. We thought that there was a neat business in building robotic products. I don’t think we felt the importance of the mission. We’re saving lives in Iraq and Afghanistan; sending the robots to Fukushima reactor to clean up a global disaster; finding the underwater pools of oil in the Gulf of Mexico that helped drive a more holistic cleanup of that ecological disaster. The impact is breathtaking. “Did you really think you were going to do that?” No, we wanted to build cool stuff. Predicting the impact on the future is quite hard, but we have a mission to continue to drive and build the industry. I will tell you, rarely is it a boring day.
X: So…what’s the coolest robot you have at home?
CA: The coolest would be Roomba. But I have this vision that these [other robots] are going to be in my home, and I’m going to do wonderfully useless things with these robots. My goal is having robots that can traverse my home, where maybe I’ll get one of Rod [Brooks]’s Rethink torsos to be a bartender, I’ll have Avas going around taking drink orders and then going back and using facial recognition to go find you and bring you your drinks. My wife will want the robot, in its circuit of the home, to make sure that the toilet seat is down.
I’m going to think of the most useless things that can be conceived of and then spend some evenings programming these cool ideas and using the tools. I think that to some degree this offers me an opportunity to get back and get my hands a little dirty. It starts from having this actual product-quality reliable navigation enabling a thousand interesting applications, from super high value—diagnose stroke from across the world—to something completely “just fun.”