Magazine

Reprogramming Amazon


It's crunch time for retailers, and Amazon.com Inc. (AMZN) is no exception. At its nine massive distribution centers from Fernley, Nev., to Bar Hersfeld, Germany, workers scurry around the clock to fill up to 1.7 million orders a day -- picking and packing merchandise, routing it onto conveyors, and shipping the boxes to every corner of the world. Like any retail warehouse running manpower and machinery at full holiday throttle, it's an impressive display. But utterly misleading. The kind of work that will truly determine Amazon's fate is happening in places like the tiny, darkened meeting room at its Seattle headquarters where, one recent afternoon, five intent faces gazed at a projection screen.

Jeffrey A. Wilke, a compact, intense senior vice-president who runs Amazon's worldwide operations and customer service, and an engineering team were trying out a "beta" or test version of new software they wrote. When the buying automation program is ready for prime time in mid-2004, Amazon's merchandise buyers will be able to chuck reams of spreadsheets for graphics-rich applications that crunch data for them, so they can more quickly and accurately forecast product demand, find the best suppliers, and more. The effort is one of scores of technology projects under way at Amazon that ultimately may change the entire experience of shopping online -- and Amazon itself.

Just as most folks have come to view Amazon as a retailer that happens to sell online, guess what? It's morphing into something new. In ways few people realize, Amazon is becoming more of a technology company -- as much Microsoft Corp. (MSFT) as Wal-Mart Stores Inc. (WMT). "What gets us up in the morning and keeps us here late at night is technology," says founder and Chief Executive Officer Jeffrey P. Bezos. "From where we sit, advanced technology is everything."

No, Amazon isn't selling its own shrink-wrapped software or leaving the retail business behind. But developing technology is becoming at least as important as selling Harry Potter books or The Strokes CDs. Indeed, some analysts say it's possible that in a few years so many other retailers will be using Amazon's tech expertise to sell on its site that they could account for more than half the products sold on Amazon.com. Says Bezos: "Amazon Services could be our most important business."

Already, Amazon's technological efforts have helped it reduce costs and boost sales so much that revenues are expected to surge 32% this year, to $5.2 billion. As a result, by the time the glittering ball descends in Times Square on New Year's Eve, Amazon may well reach a milestone some never thought it could: its first full-year profit (chart). No wonder its stock has rocketed 152% this year, to $49.34 a share.

But all that is just the start. Building on a raft of tech initiatives, from an ever-richer Web site to new search technology (table), Bezos aims to reprogram the company into something even more potent. The notion is to create a technology-driven nexus for e-commerce that's as pervasive and powerful as Microsoft's Windows operating software is in computing. That's right: Bezos hopes to create a Windows for e-commerce.

SELF-REINFORCING CYCLE

Far-fetched? Not necessarily. After all, the Amazon.com Web site is already essentially a giant application that people simply use over the Web rather than in their personal computers. And bit by bit, just as its Washington neighbor did two decades ago, Amazon is building what techies from Silicon Valley to Redmond call a platform: a stack of software on which thousands or millions of others can build businesses that in turn will bolster the platform in a self-reinforcing cycle.

Since last year, Amazon has been steadily turning innovations it developed for its own retail site into services available online. Using these so-called Amazon Web Services, reached via a browser, merchants who want to sell more can use its patented one-click purchasing system, for instance, or tap quickly into sales data for particular products. Even independent programmers are getting interested: In just 18 months, up to 35,000 programmers have downloaded software that enables them to pick and choose Amazon services and, much as they do with Windows, write new applications based on them.

Such developments recall the same creative spark that launched the PC industry and, more recently, put the Linux operating system on the map. One program makes it easy to list products for sale on Amazon. Another lets merchants instantly check prices at Amazon via a wireless Web device when they're looking at stock to buy. It's a volunteer army that costs Amazon almost nothing. "I see them not just as a place to sell things but as a provider of technology," says programmer Paul Bausch, author of Amazon Hacks: 100 Industrial-Strength Tips & Tools, a new book on how to use the technology behind Amazon's site.

Amazon doesn't make money directly from the Web services. Merchants and developers can get free access to the services and can use them to sell from any outpost on the Web. Still, many of the merchants who use applications spawned by Web services end up selling their wares to the 37 million customers assembled at Amazon.com. When that happens, Amazon takes a commission of about 15%, and these revenues have much higher margins than Amazon's own retail business. Already, 22% of Amazon's unit sales are by other merchants -- and the Web-services push hasn't had a significant impact yet.

BUSTING OUT

While Amazon isn't exiting retailing, its tech initiatives could, to some extent, set it free. By beefing up its technology and distributing it more freely, Amazon could bust out of the conceptual prison of stores and the virtual confines of a single Web site. Says e-commerce consultant and author John Hagel III: "It's really breaking apart the whole store metaphor." Into what? Already, Amazon has applied its own technology to forge an identity as an online mall -- a piece of business that generates gross margins about double its 25% retail margins. Plugging into its massive e-commerce system, thousands of retailers from mom-and-pop shops to Lands' End Inc. (S) and Circuit City Stores Inc. (CC) sell through its site. Amazon even runs the Web sites and distribution for the likes of Target Corp. (TGT) and Toys 'R' Us Inc. (TOY), which are featured on Amazon.com.

If Bezos' new plans work, Amazon could become not just a Web site but a service that would allow anyone, anywhere, to find whatever they want to buy -- and to sell whatever they want almost as easily. For Amazon, that means its finances could look considerably better than traditional retailers'. Already it turns over inventory 19 times a year, nearly double Costco Wholesale Corp.'s (COST) 11 and almost triple Wal-Mart's 7. Yet unlike most retailers, it hasn't had to trade asset efficiency for profits. Its 25% gross margin is nearly double Costco's and three points higher than Wal-Mart's.

How much Amazon can expand the narrow margins of retail remains to be seen. But if even the conservative forecasts of analysts are correct, Amazon has a lot of upside in coming years. Shawn Milne of SoundView Technology Group Inc. (SNDV) sees sales continuing to rise no less than 15% a year through 2008, to $11.3 billion, with operating margins steadily marching up from 7% this year to 11.6% in five years. That's hardly the territory of Microsoft, with its 35.5% margins, but it's more than double Wal-Mart's 5.3%. "You're going to see the company look a lot different from a retailer in terms of financial metrics," says Milne. "Opening up the technology to other sellers is what has gotten the Street interested."

IMPOSSIBLE DREAM?

Bezos' vision won't be easy to fulfill. Trying to be a world-class retailer, a leading software developer, and a service provider simultaneously strikes some observers as a nearly impossible endeavor. Even some Amazon partners report shortcomings in merchandising and technical support. Others worry about Amazon's inherent conflict in playing both retailer and mall owner. "They're biting off a lot," says Forrester Research Inc. (FORR) analyst Carrie A. Johnson. "That's their biggest risk."

Amazon is hardly alone in its ambitions, either. EBay Inc. (EBAY), the Web's largest marketplace, is building its own e-commerce platform. Already, it boasts several million sellers, at least 37 million active buyers, and more than $20 billion in gross sales -- quadruple Amazon's and is far more profitable because eBay doesn't handle goods. Increasingly, merchants of all stripes view the two companies as key channels to online customers. And eBay isn't the only contender. Search upstart Google Inc. and even Microsoft, each with its own Web services initiatives, also aim to be hubs for connecting both shoppers and merchants.

Still, Bezos' bet on technology has paid off so far. Consider what has happened in its much-criticized distribution centers, which Amazon spent $300 million to build. Back in 2000, they were eating up at least 15% of sales, partly because processes to pick and pack different items such as books, toys, and CD players weren't very efficient. Chutes holding pending orders got backed up when products didn't arrive on time. Rampant mistakes required expensive manual fixes. Software was primitive, too: Workers had to enter data into the system using arcane Unix software commands.

Now, by most accounts, the warehouses hum more like Dell Inc.'s (DELL) build-to-order factories. With a menu-driven software console, workers can anticipate where bottlenecks are likely to occur and move people around to avoid them. Another program rolled out this year sets priorities, based on current customer demand, for which products should be placed at the front of supply lines, further speeding the flow.

The result: Amazon's distribution centers can handle triple the volume of four years ago and cost half as much to operate relative to revenue, just 7% of sales. Wilke believes further software improvements can boost productivity by up to 10% a year. "They couldn't do this without very, very good software," says Stephen C. Graves, a professor of management science and engineering systems at Massachusetts Institute of Technology's Sloan School of Management who has helped analyze Amazon's distribution operations. Another bonus for Amazon: With fewer distribution mistakes, it has reduced customer service contacts per order by 50% since 1999.

All that has kicked off the virtuous cycle that's now propelling Amazon toward more consistent profits. Lower costs have enabled it to offer more product discounts and free shipping on most orders over $25. Those moves are credited with recharging sales -- and with operating costs up only 5% this year, the benefits have dropped right to the bottom line.

Even so, Bezos isn't resting easy. While Amazon's spending on technology likely will remain fairly steady next year at about $216 million, thanks to declining tech prices, Chief Technology Officer Al Vermeulen says he will hire hundreds more software engineers and computer scientists in the next year to slake Bezos' thirst for tech. "Jeff is a very big driver of the technology," says Amazon director Tom Alberg, managing director of Seattle-based Madrona Venture Group.

For one, Bezos believes there's still plenty of room for improvement on the Web site itself. To that end, he hired last year what likely is a first for Corporate America: a chief algorithms officer, Udi Manber. His mission: to develop better versions of algorithms, the complex mathematical recipes that get software to do its magic. In particular, he's creating improved algorithms for Amazon's newest tech push: search. On Oct. 23, Amazon launched Search Inside the Book, a feature that allows visitors to find any word or phrase on 35 million pages in 120,000 books -- and let them read entire pages around those keywords. In the week following, average sales growth for those books was nine percentage points higher than for books not in that database.

Manber has bigger plans yet. He now heads Amazon's first Silicon Valley outpost, a subsidiary called A9 that's charged with coming up with cutting-edge search technology. It's not just a defensive shot at search phenom Google, which is testing a shopping search engine it calls Froogle. "We need to help people get everything they need, not just a Web page," says Manber. And whatever he and his team come up with -- he's mum now -- it won't be confined to Amazon.com. A9 plans to offer search services it creates to other e-commerce sites as well.

Technologies like that, in which Amazon is reaching out beyond its own site, offer the most intriguing new opportunities -- and challenges. Consider Amazon's Merchant.com business, which takes over the entire e-commerce operations of other retailers. By all appearances, it's a success. In Toys 'R' Us' most recent quarter, in which results were dismal, the one bright spot was a 15% jump in sales at Toyrus.com, run by Amazon. Other retailers seem happy, too. "Amazon is the most sophisticated technology provider and service partner on the Internet," says Target Vice-Chairman Gerald L. Storch.

But how much that business will grow is debatable. As both retailer and mall operator, Amazon has divided loyalties. "Some companies worry about creating the next Wal-Mart that's going to take their business away down the road," says Dave Fry, CEO of Ann Arbor (Mich.)-based e-commerce consultant Fry Inc., which has helped several retailers sell on Amazon. As a result, many are going with rivals such as GSI Commerce Inc. (GSIC), which runs online operations for 51 retailers, from Linens 'n Things (LIN) to The Sports Authority (TSA). Says Forrester's Johnson: "Ultimately, Amazon will be most successful selling their own and other retailers' products on their platform."

'AN ECOSYSTEM'

That's why Web services, unfettered by logistical challenges or business model conflicts, may offer the most expansive potential of all Amazon's tech initiatives. Nowhere is the potential more apparent than among the small merchants and independent programmers who are flocking to Amazon Web Services. "It's like an ecosystem," says Bezos. "People are doing things that surprise us."

People like Cleveland Wilson, a former tech recruiter who started selling books on Amazon in early 2002. Until this year, he didn't even think it would be enough to pay his rent in Berkeley, Calif. But he discovered a program called ScoutPal that uses Amazon Web services to let booksellers type or scan in a book's ISBN number to a laptop or wireless Web device and instantly see what it's selling for on Amazon. So when he visits thrift stores and garage sales, he can buy books he knows will be profitable to resell. For example, he paid $8 for Suicide and Attempted Suicide: Methods and Consequences and sold it for $275. From $100,000 in sales this year, he expects to do $250,000 next year because he has cut the time it takes to buy and sell books by two-thirds. "Without Amazon Web services, I wouldn't be in this business," he says.

That's not all. Now, he's starting a company to develop his own software using Amazon Web services that will help other sellers improve their businesses. It's still early, but it's possible that Amazon has latched onto one of tech's juiciest dynamics -- a self-reinforcing community of supporters. Indeed, it seems to be harnessing the same "viral" nature of the open-source movement that made Linux a contender to Windows. Says Whit Andrews, an analyst at Gartner Inc.: (IT) "It creates an enormous community of people interested in making Amazon a success."

Even Microsoft. In Office 2003, people can click on a word or name in any document and be whisked off to Amazon.com so they can buy a related book or other product. Says Gytis Barzdukas, director of Microsoft's Office product management group: "It gives Amazon the ability to market to a whole new set of customers and to become a part of people's work processes."

Amazon, the next Microsoft? Not so fast. For all the promise, building a broad platform is about as tough as a goal gets in the tech business -- as nearly every competitor to Microsoft can attest. And unlike most tech companies, Amazon also has to contend not just with bits and bytes but also with the bricks and mortar of warehouses and the fickle fingers of Web shoppers. But for now, at least, Amazon's pursuit of cutting-edge technology has given it time to figure out what comes next.

By Robert D. Hof in Seattle


Cash Is for Losers
LIMITED-TIME OFFER SUBSCRIBE NOW
 
blog comments powered by Disqus