Amazon.com (AMZN) has installed more than 15,000 robots across 10 U.S. warehouses, a move that promises to cut operating costs by one-fifth and get packages out the door more quickly in the run-up to Christmas.
The orange 320-pound robots, which scoot around the floor on wheels, show how Amazon has adopted technology developed by Kiva Systems, a robotics company it bought for $775 million in 2012. Amazon showcased to reporters on Sunday ahead of Cyber Monday, the biggest online shopping day of the year.
The robots are designed to help the leading U.S. online retailer speed the time it takes to deliver items to customers and better compete with brick-and-mortar stores, where the bulk of Americans still do their shopping.
The robots also may help Amazon avoid the mishaps of last year's holiday season, when a surge of packages overwhelmed shipping and logistics company UPS (UPS) and delayed the arrival of Christmas presents around the globe. Amazon offered shipping refunds and $20 gift cards to compensate customers.
Amazon deployed the robots this summer, ahead of the key holiday quarter, when the company typically books about one-third of its annual revenue. The updated warehouses are in five states -- California, Texas, Florida, New Jersey and Washington.
The move comes at a cost. Amazon estimated in June 2013 that it would spend about $46 million to install Kiva robots at its warehouse in Ruskin, Florida, including $26.1 million for the equipment, according to company filings to local government.
The Kiva robots have allowed Amazon to hold about 50 percent more items and shorten the time it takes to offer same-day delivery in several areas, said Dave Clark, senior vice president of worldwide operations and customer services.
At Amazon's warehouse in Tracy, California, workers stack goods in shelves carried by more than 1,500 Kiva robots, which use markings on the floor to navigate and form a "big block of inventory," Clark said.
Squeezing the racks of items closely together eliminates the need for workers to navigate aisles to collect items ordered by consumers. Now, a worker calls for specific items and the robot steers itself to their particular work station. Each robot can carry as much as 720 pounds.
In some cases, the robots have allowed Amazon to get packages out the door in as little as 13 minutes from the pick stations, compared to about an hour and a half on average in older centers.
"It's certainly proving out that it's justified itself," Clark said of the Kiva acquisition. "We're happy with the economics of it."
Authored by Deepa Seetharaman via dailyfinance.com.
The Web is full of personalized content, whether it’s a Netflix recommendation or the results of a Google search.
But consumers have protested when e-commerce companies have extended their behind-the-scenes personalization to prices, charging different sums for the same goods, or pushing some people toward higher-priced offers.
A new study of top e-commerce websites found these practices—called discriminatory pricing or price steering—are much more widespread than was previously understood.
The study, by a team of computer scientists at Northeastern University, tracked searches on 16 popular e-commerce sites. Six of those sites used the pricing techniques; none of the sites alerted consumers to that fact.
Among the study’s findings: Travel-booking sites Cheaptickets and Orbitz charged some users searching hotel rates an average $12 more per night if they weren’t logged into the sites, and Travelocity charged users of Apple Inc. ’s iOS mobile operating system $15 less for hotels than other users.
Home Depot Inc. shows mobile-device users products that are roughly $100 more expensive than those offered to desktop-computer users. And Expedia and Hotels.com steer users at random to pricier products, the study said.
“In the real world, there are coupons and loyalty cards, and people are fine with that,” said Christo Wilson, an assistant professor at Northeastern who led the research team. “Here, there’s a transparency problem. The algorithms change regularly, so you don’t know if other people are getting the same results.”
Travelocity, a unit of Sabre Corp. , didn’t respond to a request for comment.
Home Depot didn’t dispute the accuracy of the findings, but the home-improvement retailer wasn’t “intentionally steering search results,” said company spokesman Stephen Holmes.
Many factors could influence what a customer sees on the company’s sites, Mr. Holmes said, including prior browsing and purchase history, the location of the store, and whether the customer is on mobile or not.
Home Depot didn’t charge users different prices for identical products but showed more-expensive products to people who shopped using a smartphone, the researchers found.
Chris Chiames, vice president of corporate affairs at Orbitz Worldwide Inc., said in an email that the company clearly advertises its loyalty programs and other deals. He said the discounts some members see on the site apply to just a small proportion of hotels—fewer than 5%. So, it wouldn’t make sense, and might even be misleading, to advertise lower prices to all members, he said.
“The Northeastern study states that ‘overall, most of the experiments do not reveal evidence of steering or discrimination,’ and so we are curious as to why a handful of exceptions to searches on thousands of hotels is the basis of this paper’s conclusions, or even worthy of a story,” Mr. Chiames added. “Would you be as interested in a Kmart ‘blue light special’ deal that was made available to shoppers who happen to be in a certain store at a certain time?”
Moreover, some deals are priced by Orbitz’s hotel partners, he added. “The hotel might have limited inventory of that price, and so they choose to display the rate on a more-limited basis, akin to the flash sale,” he said.
On Orbitz and Cheaptickets, also owned by Orbitz Worldwide, consumers who registered through the websites’ free log-in were shown a tab labeled “members only” that offered lower hotel prices. The company didn’t advertise that users could receive discounts for logging in.
Orbitz has been accused of price discrimination in the past. A 2012 Wall Street Journal investigation found the company charged Mac users as much as 30% more than PC users for a night’s lodging. The company discontinued the practice, which it characterized as a month-long experiment. The Northeastern researchers confirmed the company no longer discriminates between Mac and PC users.
Expedia and Hotels.com, both units of Expedia Inc., don’t show different prices because of users’ differing characteristics but because the company constantly refines its pricing strategies using a method called A/B testing, the researchers said. Shoppers are randomly placed in a group that highlights either less or more costly hotels. “Either way, the user has no idea what bucket they’ve been placed in,” said Northeastern’s Mr. Wilson.
One group of users, for example, were shown an average hotel listing price of $187 a night. The other group saw prices that were $17, or roughly 10%, less.
“Presenting different booking paths and options to different customers allows us to determine which features customers appreciate most” said Expedia spokesman Dave McNamee, in an email. “Pricing is not manipulated by Expedia.com.”
Consumers have long protested price discrimination. In 2000, Amazon.com Inc. Chief Executive Jeff Bezos apologized for an internal research program in which consumers were shown different prices for identical products. He called the experiment a “mistake.” (Amazon and eBay Inc. weren’t included in the Northeastern study because those companies’ services have little power over the prices they charge, the researchers said.)
Staples Inc. varied its online prices based on users’ locations, according to a 2012 article in The Wall Street Journal. The researchers didn’t examine Staples but pointed out that retailers might vary prices by region because the cost of procuring a given product can differ in different parts of the country. They didn’t study geographic variations for that reason.
In their study, the Northeastern researchers devised a statistical method to weed out what they called “noise,” or legitimate factors that might cause prices to vary—a technique they hope will be used in future studies of online personalization.
The research team recruited 300 beta users, who allowed the researchers to track their experience on different sites.
The team also developed hundreds of fake accounts to see whether historical purchase patterns and clicks through the sites had an impact on price personalization. They didn’t examine the impact of consumers’ overall Web-browsing behavior on pricing because they would have no way to know how e-commerce sites tracked participants across the Web.
Write to Elizabeth Dwoskin at firstname.lastname@example.org