Several months ago Amazon bought Kiva, a firm that develops robotic fulfillment systems. Several photos and tours of Amazon’s newest warehouses were released this week, and people were shocked to see that the warehouses are still built for human pickers, with no robots in sights. Bloomberg’s BusinessWeek ran an article trying to explain this observation. (Amazon’s Robotic Future: A Work in Progress)
After all, aren’t robots supposed to be the future of such places as distribution centers and warehouses? Didn’t Amazon buy a robot manufacturer, Kiva, in March? The online retailer announced in October that it was taking on 50,000 additional part-time workers for the holiday season. Shouldn’t some of those spots be taken up by mechanical arms and wheels?
The article provides several explanations:
Bruce Welty is chief executive officer of Quiet Logistics, an order-fulfillment company that manages the online inventory and distribution for retailers like Gilt, Zara, and Bonobos. He uses robots made by Kiva, the company Amazon purchased, but his warehouse in Massachusetts is not bereft of humans. “Robots aren’t very good at picking up things,” he says. “They aren’t very good at looking at a bin of different things and distinguishing one item from another.
I am not sure that indeed robots aren’t good at picking up things. In my opinion, the main issue is that robots are not good at picking a variety of things at different sizes and shapes. Given Amazon’s variety, this lack of flexibility is problematic. A second reason is also related to flexibility. In this case, volume flexibility:
Automation won’t help Amazon in periods of peak demand,” says Stephen Graves, a professor of management science at the Massachusetts Institute of Technology. “People are far more flexible.
This is an explanation we already provided in one of our earlier posts about Amazon . Given that during the holiday season the demand goes up, sometimes fourfolds, the most cost efficient way to do it is using casual employees. The capital expense to add robots means they need to make sense throughout the year, and not only during the holiday season. The third explanation provided as to why Kiva was purchased, is that it was for the software, as much as the hardware.
Of course, there is another explanation, which is that Amazon is now as much about books and music, as it is about appliances and safes. Furthermore, Amazon may have a role for these Kiva robots that is not necessarily in the “last mile” of the “pick and pack” operations, but in the intermediary steps. While clearly the entire system has to ramp up during the peak season, I expect the “pick and pack” operations to be more exposed to these fluctuations. Other segments may be easier to smooth, and thus are more suitable for robot operations.
That’s a fascinating question. I’m wondering if maybe this is partly risk management on Amazon’s part: if they make a big change to their packaging process a few months before the holidays, something is more likely to go wrong than if they make the transition incrementally. I guess the unanswered question is, “How expensive is it to retrofit the factories for robot workers?” If it’s cheap, then it might make sense to wait a year; if it’s expensive, then maybe not.
Will,
That’s definitely a plausible explanation, and I am sure it has something to do with the current status. Yet, the purchase was done several months ago, and the new warehouses could have been prepared for that. This is not the case.