iRobot CEO talks about the Market, the Competition and the Future

  Robot vacuuming is becoming globally “mainstream,” says Colin Angle, Co-Founder and CEO of iRobot, the maker of the robotic vacuum cleaner Roomba. It’s been 8 years since the product was first introduced into the market, and this year sales have taken off outside the U.S. The annual worldwide market for vacuum cleaners that cost more than    $ 200 is 4 billion dollars and iRobot plans to capture about 5% of that market in 2010.
 So what’s next and what about the competition? GetRobo interviewed the CEO on the phone to talk about what he thinks. (This interview was conducted originally for a column on the Wall Street Journal Japan.The following is an edited transcription of the conversation.) 

Q. You have sold over 200,000 Roombas in Japan and I hear that it's gaining even more momentum there. What’s happening on the Home Robot front ?
 
A. This is an emerging market place. The home robot industry is still a relatively new industry and we’re enjoying a period of time when people are starting to understand that this is not just science fiction and that robots are actually practical and reliable. I think it’s fair to say there still is this skepticism that people have around the reality of robots in general. But since we launched originally in 2002, we’ve sold about 5 million (units) worldwide and I think we’re really at a point where we’re seeing a “mainstreaming” of robot vacuuming.
Our international strategy has been very successful in getting that message out  and Japan is our largest single international customer. 
 
Q. You talk about your IP strategy and your “strong defensible intellectual property portfolio” in your most recent earnings conference call. How are you going to use this portfolio and how do you view the competition right now?
 
A. We have and we’ll continue to pursue active defense of our intellectual property. It’s not something you do lightly. Typically the way this is done is because of the investment required to take action, you wait for a competing product to demonstrate that it has some staying power in the marketplace. Since we launched the product in 2002 there has been at least 20 or 30 attempts to copy in varying degrees, the functionality, design and methodology that Roomba utilizes. And typically what happens with these products is they fail in the marketplace on their own because frankly building a practical robot that effectively cleans the floor is an extremely difficult thing to do.
 And then you get some other companies that have an innovative element to the solution but have not solved enough of the problems to have a viable product. So, here’s a cool feature but it can’t clean. 
 This is something we track very carefully. We certainly are seeing some products that seem like they have some serious commitment behind them. But we don’t think there is anything in the marketplace that comes close to Roomba in solving the real problem, which is automatically vacuuming your floor. There are a thousand different little things that competing companies need to get right and it will be some time before frankly there is legitimate competition in the space. And when that happens, that’s not necessarily a bad thing either, because certainly the vacuuming industry is quite large and the more quality products that are out there helping people understand that robot vacuuming is the better way of cleaning your floors, the more rapidly we will see more adoption.
 
Q. So far, have you used your IP portfolio for legal actions in the home robot market?
 
A. We have. We took action against a company called Koolatron that was importing a product into the U.S. a few years ago that we started to see in the channels that we were at. We took action and got a consent judgment against that company and forced them to cease and desist. We’ve done it before and we’re certainly willing to do it again. We’re tracking this new raft in entries very carefully to make a prudent decision as to timing and the identity of the folks.
 
Q. How serious is the Neato XV-11 as competition?
 
A. That would be a product that has some interesting technology but not a complete solution. It has a laser range finder on it which is interesting, but it adds a lot of cost and will treat everything like a wall. So if you had a table cloth on your table or you had a couch that had skirts on it, the robot would never go underneath and that tends to be a place that you want your robot to go. So that’s a limitation. But most importantly I think they put most of their energy into the navigation and not enough energy into the cleaning efficacy. Frankly the navigation is the fun part but the rubber meets the road when you ask if it actually picks up the particular that’s on the floor.
 But it’s a good company and what I would hope is that they have enough success with this early version that they can solve more of the problems in their next generation. I’m all for startups in this space and I’m outspoken trying to encourage companies like this to enter. I’d be happier if people were trying to solve any one of the (other) thousands of challenges in areas that robots can help around the home or business and industry today, but if vacuuming is where they want to play, vacuuming is where they are going to play.
 
Q. There’s a lot of excitement around Kinect. I just saw a YouTube video about Kinect and iRobot Create. Do you feel that this kind of technology can be combined with the Roomba in the future?
 
A. Absolutely. The type of technology the Kinect sensor uses is a technology that has been used on robots for decades. When I was at MIT we had a point cloud sensor but it was very very expensive. What the Kinect sensor gives us is a low cost way in creating a 3D representation of our environment. That’s something that robots can use to do a more sophisticated job of navigation or identification of what the environment contains.
 Or you can simply use some of the software that’s being developed for the Xbox where gesture is translated into meaning and so you can point at a place you might want the robot to clean and let the robot do it. I think it opens up a style of interface which used to be unaffordable in a consumer robot. We actually have gestural interfaces created at the research level for military robots so we are very aware of what can be done. And software and computation are not the limiting factor anymore - thanks to Moore’s Law - it’s having the right sensor. The Kinect sensor is a great one and we’re really excited about its potential. 
 
Q. Does iRobot plan on having some capability for 3D representation of the environment on the Roombas?
 

A. I would say that certainly it is being evaluated at the research and development stage. It takes a real amount of time before you want to integrate that into a robot system. So it’s not something that you should expect to see next year but it’s an exciting technology and if the price of it delivers enough value to the customer, then something like that could very well be seen.
 The Kinect retails for the Xbox at $150 in the U.S. and that’s just for the sensor. So the question is, you have a $300 dollar robot and how many people want to spend another $150 in order to be able to have a gestural interface in the robot. These are the kinds of tradeoffs that you have to make and if the market says, hey, that would be worthwhile, eventually you’ll certainly see something like that. But I’m not foreshadowing a decision or a timeframe. You would be right if you wondered whether or not the gestural interface to a Roomba really warranted 50 % increase in its cost. There may be other robots and other applications of robotics which might be more suitable for that kind of interface. But this is a very important sensor for the robotic industry because it opens up the opportunity for robots to do more sophisticated things than they were able to do before.
 
Q. I was thinking more about something like the Kinect being able to offer the capability of the laser range finder on top of the Neato XV-11….
 
A. Potentially yes, but again you get the same cost question. Roomba currently extremely thoroughly covers your floor. We don’t have a navigation problem or limitation. It basically uses a series of algorithms to ensure coverage without relying on sensors which would be easily spoofed by skirts around couches or beds and things that would block the robot from going into the very places that should be vacuumed.
 Roomba is more algorithmic and heuristic in how it approaches things. It’s absolutely not random like people say. It has strategy depending on where it is. We have technology like the Neato system but our concern is that it cleans less well when you do it that way.
 So looking at the Kinetic sensor, you say OK, what is the customer benefit and does that merit the additional cost. If it looks smarter and it cleans worse, and it’s more expensive, that’s not a winning proposition for us.
 
Q. In the past, you have talked about working on robots for the elderly. Can you talk about what kinds of robots you are working on and how you see the future of this market?
 
A. At this point we’re not talking about the specific products that we are working on, so I apologize but this is relatively early stage for us. We view this as a very important long term opportunity for robotics but again we are not announcing, foreshadowing, or setting any expectations that you should see a product eminently in this space.
 We’ve talked about it because we think it’s important. We are working in this space and (since we are) working with partners we thought it best to be open about it rather than closed. I think in an emerging market place like robotics, at least our company’s philosophy is, as long as you’re not setting any expectations and promising products, (you could be) relatively open about what you are doing. It facilitates the industry to grow. So our motivation there was to say we believe in this, if you’re interested come talk to us.
 Now, I am happy to talk a lot about why I think this is an important area. I The impending crisis of care - it’s even more acute in Japan than it is in the U.S. - and it’s only going to get more so. The idea of living in assisted living facilities or nursing homes is just not economically feasible for virtually everyone that’s not extremely wealthy unless the government is paying for the vast majority of it. And the government is increasingly not able to take on that burden. So there is a huge need for either the society changes how it views care of the elderly or technology comes into existence that allows people to live independently. We think that latter need creates a huge economic opportunity. 
 When you think why people need to live in assisted living, their physical issues lend themselves to robotic solutions whether it be getting up in the morning, getting dressed, going to the bathroom in the middle of the night, etc. There’s a wide array of applications where mobility and manipulation and physicality are going to be incredibly important.
 
Q. For the first time, iRobot appointed a COO.
 
A. Yes. We got into a scale where in order to manage our long term strategy as well as increasingly complicated divisional organization I decided that the company would be better served if I had a COO to help me divide my time more effectively.
 
Q. You mentioned in the conference call that the new COO will allow you to spend more time focusing on long term strategy, pursing M&A and alliances. What kind of M&A and alliances will you be looking for?
 
A. The challenge with robots is it embodies so many different technologies that it would be naïve to think - and it wouldn’t be a successful strategy - to try to lead the world in all these different areas. So iRobot by focusing on artificial intelligence and Aware2 software, core robot competencies, we can extend our current leadership positions while looking to, for example , the mobile computing industry for things like voice recognition, connectivity and wireless communication. We are a consumer of those technologies and need to have partners in those areas. It creates a need for outreach and teaming.
 As far as M&A goes, we’d be looking at component technologies that are primarily utilized in the robot industry. We talked a lot about the Kinect sensor. It would be interesting if we owned that. Well, the opportunity in robots is relatively immature. This is one of the hit products for Christmas 2010 and all of the economic benefit will lie in the video game industry. That is less interesting to us because we aren’t a video game company. We’re a robot company so rather if we found companies that had technology like the ASC flash LADAR which is a outdoor ranging sensor - and similar to the Kinect sensor - whose primary utility might be in the robotics industry, we’d be more interested in investing in that. (Writer note: In 2007 iRobot entered into agreement with Advanced Scientific Concepts, Inc. for exclusive rights to use its LADAR technology for unmanned ground vehicles in exchange for future commitments to purchase units.) Companies that have a cool robot product which we thought we could put through our distribution channels would be a great interest to us too.

 Q. Your company has a “hybrid” business model with 2 divisions (Home Robots and Government & Industrial). What are the advantages of having these 2 divisions within one company?
 
A. There are a number of different advantages. First off, in an emerging marketplace where you are really trying to change or bring new things to market, diversity does help. There have been times when the characteristics of the military marketplace provided the stability the company needed to maintain the scale that we now enjoy. During 2009 when we were in the throes of a recession, we had strong support and strong performance in our military side which certainly helped. And there are times when the rapid growth in our home robot division gave us the importance. So diversity is an important element.
 Another element is the fact that these two industries are synergistic and we see technology transfer (between them). Expertise on the consumer side means having to work at low price points. Also we have to think about millions of quantities and reliability in millions of households in all different types. It puts a lot of pressure on what sorts of technology that we can develop.
 On the military side, it’s lower volume and more specialized, which gives you an advantage where you can spend more money on a particular component and gain experience with that component and then in the future look for ways to make that particular robotic element cheaper.
 As a result (of having 2 divisions), our military products are more cost effective because we are able to think about and have experience in manufacturing techniques that companies without a consumer side would have less ability to execute.
 On the consumer side, we have this amazing technology feeder where algorithms and artificial intelligence software and sensors that would be too costly to be incorporated into a home robot are developed and perfected for the government customer. So it’s neat how it works. 
 Additionally the government model allows us access to government sponsored research and development dollars so that we can support very large blue-sky robot technology (development). The way (the government sponsored R&D projects) work is that the U.S. government retains government use rights but the performing company maintains all commercial rights to the technology. Anything we develop under government funding can be commercialized with full IP protection throughout our non-government businesses.
 
Q.  Are the R&D departments of the 2 divisions together or separate?
 
A. The way we structured it is that we have a centralized research and development group led by our CTO that works on the cross-cutting technology and research. And then in the divisions we have advanced development and product development. Product development is when you know what you want to do and you go develop a product. Advanced development is when you think you know what you’re trying to do and you are working on the advanced technology so that they can be applicable in a product.
 So for example, if we want to make a tree-pruning robot, first we need to learn how to climb trees. Advanced development would be to take academic tree climbing technology and reduce it to practice so that you know that you can climb trees. And then, the product development team would say OK, now we know it can climb trees and we know how it’s going to pick fruit, we’re going to build a product that pulls them all together.
 
Q. As CEO of iRobot, what is your number one challenge in the next 2-3 years?
 
A. I think in the next 2-3 years, we’re going to continue to see new products come out and existing products work to achieve success with a growing constituency. I think that robotics still is an early stage industry and the skepticism around how people should think about this industry is still very high. So I view over the next 2-3 years, further legitimization of the industry. There are economic successes in vacuuming, in military surveillance robots, and some success in warehouse automation type of tasks – companies like Kiva Systems. But I think we need more and bigger successes for the industry to really grow beyond interesting. We need more economic successes.
 Over the next few years we’re going to see that happen and that will be very exciting for me personally as a long-term proponent and innovator in this space to see this industy continuing to take off. And iRobot in particular to watch how we can grow our leadership position in face of a increasingly interesting backdrop of companies.