Don't call them robots.
I made the mistake five minutes into a conversation with Kiva Systems Inc. chief executive Mick Mountz, who paused for a second, smiled, and explained:
"We're not about the robot. This company was founded to solve a business problem, and a lot of robotics companies are about a cool technology that is looking for an application."
The idea of turning a warehouse over to stout orange robots the size of an ottoman might induce anxiety in the average logistics manager, and so Mountz prefers to call the rolling machines built by his Woburn-based company "mobile drive units" (though he sometimes relents and uses the term "bots").
The theory is that the bots can make order fulfillment faster and more efficient, letting a human stuff more boxes per hour. Kiva's bots can also rearrange warehouses on the fly, moving racks of popular items closer to the workers, while consigning slow-selling items to a distant corner. So far, customers like Staples Inc., Walgreen Co., and the online shoe store Zappos.com have been willing to give Kiva a try.
Kiva is part of a growing cluster of Massachusetts companies that are developing a new generation of robots that can do surprising things: clean out rain gutters, swim underwater to inspect the hulls of Navy vessels, and manage warehouses. The state has more than 150 companies and research labs working on robots, according to the Massachusetts Technology Leadership Council, which says the figure is conservative.
"Boston and Pittsburgh both have a good number of robotics companies, but Boston has the advantage, in terms of being a larger city, and a larger investment community, " says Dan Kara, president of Robotics Trends, a Natick company that organizes the annual RoboBusiness conference; it alternates between Massachusetts and Pennsylvania. Silicon Valley, Kara says, has a decent number of robotics companies, but doesn't register very high on the robotics Richter scale.
Kiva's roots are in the Valley. Mountz worked briefly for WebVan Group Inc., an ambitious online grocery delivery service that raised $375 million in an initial public offering, and at one point employed 2,000 people. Mountz says that the high cost of fulfilling orders was one of the factors that sank WebVan, which declared bankruptcy in 2001. He started thinking about better approaches that would combine intelligent software and sturdy bots, and he found some engineers in New England willing to help build prototypes of the devices.
Mountz also found investors more receptive on the East Coast. "I traveled Sand Hill Road for a year, and just got a lot of blank looks," says Mountz, referring to the main drag along which many Silicon Valley investors maintain offices. Through a business school connection, Mountz was introduced to Ajay Agarwal at Bain Capital Ventures. Bain not only became Kiva's main investor (the company has raised $18 million so far, from Bain and several angel investors), but also introduced Kiva to executives at Staples, another company that Bain initially backed.
As a part owner of companies like Toys "R" Us, Bain "understood the logistics issues that companies face," Mountz says. He moved east in January 2004. In most warehouses, employees either walk the aisles and pull products off shelves, placing them in the industrial equivalent of a shopping cart, or they stand in place and fill boxes that roll past them on a conveyor belt, reaching for products on shelves nearby. Kiva's approach is different: The robots bring products to a stationary worker, creating what Mountz calls "an infinite shelf."
Standing at a "picking station" inside the simulated 10,000-square-foot warehouse at Kiva's headquarters, he shows how it works. He asks me to order a few grocery items on his laptop from a demonstration Web page - we choose Brigham's vanilla ice cream and Quaker oatmeal. Instantly, a printer spits out a sticker, which goes onto an empty cardboard box.
Within two minutes, Kiva's "mobile drive units" have located the racks holding those products, lifted them off the floor, and brought them over to Mountz. A laser pointer indicates which slot on the rack holds each product, so Mountz doesn't have to hunt for it. When he scans the ice cream container's bar code, indicating that he has removed it from the first rack and is about to put it in the box, that robot peels off, and the next rack moves into position.
The robots navigate by looking for bar codes affixed to the floor every few feet and coordinate their movements by communicating over a WiFi network with a computer that serves as a kind of air traffic controller, preventing collisions. They can also recharge themselves periodically at charging stations.
In a typical warehouse, Mountz says that each worker is assisted by about 10 robots. A swarm of about 500 bots work at a Staples distribution facility in Pennsylvania, picking up shelving units full of products and ferrying them over to workers who are filling customer orders. Several of the bots often line up, giving the worker access to a steady stream of items to fill the box - no roaming the aisles required.
At Staples, vice president Roger Will says, the system helped improve the company's "perfect order metric": the number of orders shipped to customers on time, and containing exactly the items ordered. Staples declined to be more specific about the improvements, but Will did say that the productivity of the "picking" process - getting products off shelves and into boxes - has doubled.
Kiva executives say an entry-level system starts at about $1 million (a system includes the bots, the necessary software, moveable racks, and employee work stations), but most installations so far have been in the $4 million to $6 million price range. The company's latest customer, Zappos.com, is deploying a Kiva system to expand a warehouse in Kentucky.
Kiva could eventually face competition from other robotics companies developing new technologies for the warehouse. One, Pittsburgh-based Seegrid Corp., introduced a "vision-guided tow tractor" in 2006 that can move forklift pallets around a factory or warehouse after a bit of training from a human. Another, Watertown-based CasePick Systems LLC, hasn't revealed much about the warehouse automation system it is working on. It has 11 employees, compared with Kiva's 75, and is funded by a strategic partner rather than venture capitalists.
But like most entrepreneurs trying to bring a new technology to market, Mountz says his biggest competition right now comes from already established warehouse automation companies, eager to sow doubts about his offering.
Bringing in the bots entails a psychological rethink of how warehouse work gets done, and some prospective customers may worry about reliability and maintenance costs (though Staples says those haven't been an issue). Which is why everyone involved with Kiva would rather de-emphasize the word robot.
"If you're not dealing with the senior levels at a prospective customer, this kind of change can seem too cutting-edge," says Agarwal at Bain Capital Ventures. "We're an order-fulfillment solution that just happens to use mobile robots," Mountz says.
OK. Got it.
Scott Kirsner can be reached at firstname.lastname@example.org.