MM: Instead of having humans picking orders in the zones, our innovation was to have the products come to the worker in a pick station. The innovation was more a business process innovation - the robot drives over, picks up a whole shelf and brings it to the human order picker.
We have multiple cooperating robots, loosely defined. When we put a Kiva system into a factory, it involves from 100 to 150 robots on the low end, to 400 to 500 on the high end. Someday, we envision large warehouse installations with thousands of robots running around.
RT: How does it work?
MM: The value we bring to the customer is increased productivity and thereby cost reduction. If they have 100 people filling orders today with conveyers or other automation, we can fill the same orders with 50 people. So they get from two times to three times, and in some cases, four times improvement in productivity in their warehouses. We are doing some designs where the original building used 600 employees, and now they can do the same work with 150 people.
Our product is a solution that involves hundreds and hundreds of mobile robots, and workstations where the operators interface with the product. There are many aspects to the solution.
Our product is actually a solution, so we are a little different from classic robotics companies. Kiva was not founded as a robotics company; it was founded to solve the pick, pack and ship order fulfillment problem. It just so happened that robots were a good way to solve that problem. That is different from the classic robot industry where an innovation is looking for a problem to solve. And that is what has enabled us to move very quickly. If you have a very focused problem, you can build a very dedicated robot to solve it, and have more development and less research.
RT: How long did it take to develop and put the system on the market?
MM: We took an iterative, staged approach to developing the technology. In 2004, we built a proof of concept, a small scale system, that we could show customers. We did two pilot customers that year. In 2005, we did two tests of production units of the E series robot. That allowed us to demonstrate success in a real world factory environment. At that time we were also working on the F series robot, which came out in 2006. So in three years, we went through three iterations of the platform. We have been shipping since 2006.
Kiva Systems is operating in seven states now and we have shipped over 1,000 robots. We do not disclose the actual number, but we passed the 1,000 robots shipped mark last year. We are off to the races with this commercial application of mobile robotics. Our customers include Staples, which has converted two distribution centers, Walgreen’s, Zappos.com and recently, Diapers.com and the Gap.
RT: How smart are the robots?
MM: They know how to not bump into each other. There are control systems at all levels on and off the robot managing coordination. We are solving, to some degree, the problem of complex interactions.
The robots are a highly visible component of our systems, but our solution involves everything from the concrete floor and how we navigate around it, to the software in the robot, the software in the workstations and the humans interacting with the system.
It is an inventory management, picking and planning system. We have developed a very large enterprise level software package that runs on clustered servers that direct all the activity and also talks to the customers host system. We take orders form the host system.
RT: How much does the Kiva system cost?
MM: On the low end, it runs from $1 million to $2 million. Our sweet spot is the $4 million to $6 million system. We have done some $10 million installations. It depends on how much work the warehouse is trying to accomplish, in orders per day or orders per hour.
The Kiva approach, which is novel in the material handling world, is to make very low cost robots so customers could have hundreds of them. That is the opposite of building expensive, automated guided vehicles. One of those might cost $150,000 and a company might put in 10.
We have low-cost robots so customers can put in hundreds. A $1 million system might use 30 to 50 robots. Kiva robots are to automated guided vehicles what the PC was to the mainframe.
RT: Who are your competitors?
MM: We have a lot of competition. Companies setting up a warehouse to do order fulfillment can choose from a dozen different approaches - manual picking, pick carts, vonveyers, automated storage and retrieval machines, things like that. But there is nothing like the class of product that Kiva provides. There is no direct robotic competition for us.
RT: What are the main challenges you face?
MM: We are a high-tech startup growing rather rapidly, so we have the classic challenges of finding the right employees, and making sure we get the correct installed base of customers. We need a growing infrastructure.
RT: Why did you decide to locate in the Boston area?
MM: The Boston area is a great market to be in robotics. There are hardware and firmware companies in the area, and great software people. We are finding the people we need.
And one of the hidden benefits of being a robotics company in Massachusetts is that we have access to many electromechanical vendors located in the area. We actually manufacture and assemble the robots at the Kiva main office in Woburn, Mass. So we leverage hundreds if not thousands of vendors in the Massachusetts area. The industrial base is good in part because of the defense industry located here.
Building Kiva robots is a small batch, high-precision task. Many of these companies have been in business for many years. The manufacturers who have stayed and not moved overseas tend to be the more specialized, high-precision, low-volume companies.
RT: Can you comment on what is on the drawing boards for Kiva?
MM: It about doing more with the same formula. We have solved the specific problem we set out to solve. For us now, it is about further market penetration of the solution. We expanded the product line in the past year. The original robot carried 1,000 pound shelving units around. We added a model last year that carries a 3,000 pound pallet. With that, we can cover all the functions and workloads in the warehouse. Our goal is to penetrate more warehouses.