Just a few weeks back, 60 Minutes did a piece entitled "Are Robots Hurting Job Growth?" The answer was a resounding yes. If the report is to be believed, the American worker is beginning to look like the Tin Woodman in L. Frank Baum's original The Wonderful Wizard of Oz. Enchanted by the Wicked Witch of the East, he chops off his own limbs one by one, replacing them with metal prosthetics, until he's made entirely of tin. (He does end up with a human heart, something that is apparently lacking in the CEOs who shipped those jobs off to Asia in the first place.)

The 60 Minutes report was based in part on the work of MIT economists Erik Brynjolfsson and Andrew McAfee, whose book Racing Against the Machine lays out the whole scary scenario. They blame the digital revolution for persistent unemployment and stagnant wages, as human workers find themselves replaced by machines that can do their jobs faster, better and cheaper.

Brynjolfsson and McAfee have their critics. At the recent Automate 2013 trade show in Chicago, Henrik I. Christensen, who holds the Kuka Chair of Robotics at Georgia Tech's College of Computing, argued that while automation eliminates some jobs, it creates many others that pay more and require more skills. And what's so new about that, anyway? Haven't we repeatedly witnessed this kind of transformation since the dawn of the Industrial Age?

What's more, robots aren't suited for every task on a plant floor, or even a distribution center. Certainly we've seen the emergence of the "lights-out" warehouse, where picking and putaway are done exclusively by machines and there's scarcely a human to be seen. But these are highly specialized systems that can't be applied to just any type of product or building.

Even in areas that seem ripe for automation, a living, breathing person might be the better choice. Doug Sternberg is executive vice president of strategy with Dotcom Distribution, a provider of fulfillment and distribution services for both traditional and e-commerce channels. He acknowledges the "buzz" caused by operations such as Amazon.com, which bought Kiva Systems Inc., a seller of highly automated material-handling technology, in 2012 for a reported $775m.

Perhaps Kiva got on Amazon's radar in the first place because of the latter's earlier purchase of Zappos, the popular online merchandiser of shoes and clothing. Zappos had installed Kiva technology in its Shepherdsville, Ky., distribution facility. But when the company added clothing items to the site, it pulled out the machines and reinstated a manual picking operation.

That's precisely the kind of scenario that prompts Sternberg to doubt the applicability of automation to many distribution environments. (And while we're at it, let's be careful to distinguish between process automation and the more specific category of robotics. The two are often confused.) "For us in the third-party fulfillment space, automation is a very tricky topic," Sternberg says. "It's more about having the right amount rather than being bleeding-edge. We have to be very good at being flexible."

Sophisticated material-handling systems tend to be purpose-built. When a distribution center's needs evolve, either in terms of volume or product assortment, they can quickly become obsolete. Meanwhile, the lowly human stands ready to adapt to any conceivable change.

Making machines pay can be especially tricky in the fashion and apparel space, one of Dotcom Distribution's specialties. Say a client has been selling hanging garments or items shipped in flat packs. Suddenly it sees the opportunity to expand into handbags, which are big, bulky and come in endless shapes and sizes. Or the D.C. takes on a client who requires the drop-shipping of product in branded boxes with new and exacting specifications. An existing automated system might not be up to the job.

In addition, there are certain tasks that don't lend themselves to automation, at least in its current form. Product customization - including packaging and presentation - is becoming increasingly important to retailers, Sternberg says. A client might call for the addition of tissue paper, stickers, ribbons or other decorative items. Or it might deal in high-value items, such as jewelry and artwork, that require delicate handling. "You can't just put it into a tote or conveyor," he says.

Finally, there are those industry sectors that are coping with the peaks and valleys of seasonality. Asks Sternberg: If you're putting in an expensive automated material-handling system, what do you build for? The periods of highest demand? Some arbitrary average? Either way, you're looking at millions of dollars' worth of equipment that's going to sit idle for most of the year. A more manual operation can add or subtract people and pack stations as required.

"If you're locked into an automated solution, you can't quickly expand those capabilities, especially when you're in the heat of battle," says Sternberg. Never mind those areas of retail where demand is difficult, if not impossible, to forecast accurately. "It's Christmas here every day," says Sternberg of his own operation. "You never know what to expect."

That said, he isn't blindly opposed to the use of robotics or other types of cutting-edge automation in the warehouse. It's all about what makes sense for that particular operation. "You're always going to have to find the balance between flexibility and the automation used to support it," he says.

Of course, the future could bring us a breed of robot able to handle the most delicate task, and evolve in ways that we can't even imagine today. One day, we could be looking at distribution facilities full of Tin Woodmen, minus the beating hearts.

"I would be foolish to argue that," says Sternberg.

Comment on This Article


Keywords: supply chain, supply chain management, inventory control, warehouse management systems, supply chain systems, WMS, 3PL, third-party logistics, retail supply chain