Case Study: RosettaNet

By Patrick Porter

Case Study: RosettaNet



Four years ago, leading electronics parts contract manufacturer Ingram Micro Inc. and a handful of other big companies in the $136 billion contract manufacturing industry saw the writing on the wall: Get leaner and more cost-competitive, or lose business from big customers like Compaq Computer Corp. and IBM Corp. But with razor-thin profit margins—typically around 1 percent or 2 percent—and with Dell Computer Corp. turning over inventories at nearly six times the speed of the rest of the industry, everyone in the electronics supply chain had to get smarter and faster, or lose millions.

So an elite group of IT executives, including Andrew Grove of Intel Corp., put their heads together and, in an unusual, global experiment in collaboration, embarked on a grand gamble called RosettaNet. Named for the Rosetta Stone, which cracked the code of hieroglyphics in 1799, the goal of RosettaNet is to crack the computer language barriers of the Information Age, and make it easier for companies and their parts suppliers to communicate with one another electronically. The hoped-for payoff of the effort: cost-cutting and reduced parts inventories, from a bloated industry average of 65 days' worth of parts on hand down to a more cost-efficient 10-day supply, such as that held by Dell Computer, the most wired company in the chain. But the reality of RosettaNet so far? It's off the mark. Stymied by cultural and technical problems, RosettaNet is still struggling to get off the ground.

It all began ambitiously enough. When RosettaNet first launched in May 1998 as a consortium of some 40 companies, its backers predicted confidently that its virtual supply-chain technology would slash waste, speed product to market and reduce costly inventories that Stanford University supply-chain expert Haim Mendelson estimates have been, in effect, giving Dell a cost advantage of $300 million a year over the rest of the industry. Down the road, backers said, RosettaNet would also allow companies to design products collaboratively and work together using the Web to share demand forecasts with suppliers and adjust production schedules on the fly. RosettaNet co-founder Fadi Chehadé boasted in his April 1999 announcement of the project that "the IT supply chain stands to reap significant savings…possibly as high as $25 billion annually" by using RosettaNet and its cross-platform XML standard to "translate" electronic communications among thousands of companies in the chain. "RosettaNet will be revolutionary," Chehadé maintained.

While some companies have benefited so far—Avnet, Inc., for example, expects to save $2 million over the next five years by implementing some RosettaNet business process standards—few companies have followed suit. Today, only a handful of companies have begun to convert even a part of their systems to this virtual supply chain, much less reap any payoffs. Bottom line: Not all companies are ready, culture-wise, to accept the changes. And the sheer cost and complexity of getting a $1 trillion industry group to agree on a vast array of common business and IT standards are keeping RosettaNet from becoming the lingua franca of the industry's supply-chain network. While 57 business processes have been standardized so far under RN, for example, another 50 to 80 such processes still need to be agreed upon before they can be translated into a digital language that everyone in the supply chain can understand. "To ask multiple players in an industry that's facing hard times to join RosettaNet is exceedingly worthwhile but exceedingly difficult for some companies," says Mendelson. "It's asking companies to do something that benefits the industry, and not just themselves. That's a new concept for a lot of people."

Net Gains

Net Gains

So why keep pushing RosettaNet at all? The biggest reason, say industry experts, is the tremendous cost savings to be gained from harnessing systems to the Web. RosettaNet CEO Jennifer Hamilton estimates those savings could add up to between 2 percent and 10 percent of a company's revenue—big numbers considering that the more than 400 members of RosettaNet have a combined revenue of more than $1 trillion.

Take, for example, Milpitas, Calif.-based Solectron, a contract manufacturer of electronic goods. Solectron buys components from more than 7,000 suppliers to feed more than 60 manufacturing facilities where it makes personal computers, laptops, servers, cell phones, communications switches and other electronics gear for 300 leading computer makers, including IBM, Compaq, Cisco Systems Inc. and Hewlett-Packard Co. The company owes most of its 45 percent annual revenue growth over the past few years to its clients' penchant for outsourcing. It buys factories from clients in return for long-term contracts to provide those clients the same finished goods they once made themselves. One problem: Whenever Solectron acquires a customer's manufacturing operations, it also inherits existing relationships with its entire supply base, often made up of scores of companies, each with its own set of business processes and IT systems. A failure of RosettaNet to expand and take hold would slow down efforts by Solectron and other companies to get these systems to talk to each other—and deal a critical blow to future efforts by Solectron and, indeed, other industries, to standardize their systems around the Net and XML in the race for future market share.

How inefficient is today's supply chain? Every extra day of inventory in the channel costs the $136 billion contract manufacturing industry up to $175 million, says J. Keith Dunne, senior analyst at Robertson Stephens Inc. "If you could cut 20 days, that's $3.5 billion in annual savings you could use to reduce debt or expand investment," he says. For Solectron alone, the cost of holding inventory is about $50 million a day, Dunne says. Solectron has some 100 days of sales in inventory—roughly twice what is considered cost-efficient for the industry—at a cost of some $2 billion in excess inventory, Dunne says. "RosettaNet could help Solectron to cut that significantly."

Another incentive: If RosettaNet can be made to work, it might also lower IT and process development costs. "You can implement RosettaNet once and extend it across all of your business partners," says Mark Jenkins, director of e-business systems and services at Solectron. "In today's environment, speed wins out over customization both inside the corporation and across the supply chain."

Yet thanks to RN's delays in getting off the ground, Solectron and other companies have had to develop alternative Web systems to handle everyday activities, such as buying materials and quoting prices, to compensate for what RosettaNet has not yet been able to deliver: easy-to-use standards for streamlining the way electronics parts are bought and sold.

The parts supply chain is also outdated. Ironically, except for expensive, decades-old electronic data interchange systems (EDI)—by which company A blasts company B an order for 5,000 capacitors, and company B blasts in reply that they're coming in three weeks and for how much—most of the world's electronic components are still bought and sold the old-fashioned way: by hordes of clerks using telephones and faxes, jotting records on post-it notes and desk calendars.

Language Lessons

Language Lessons

Other industries are similarly IT-challenged. Whether hotel supplies or ready-made doors, dozens of industries are still populated by hundreds, if not thousands, of small specialty suppliers who still can't afford EDI, let alone the more sophisticated ERP systems needed to take humans out of the loop. "RosettaNet, with its data definition standards and low-cost Web platform, promises an innovative solution to the problem of having a highly fragmented supply base," says Stanford University supply-chain expert Hau L. Lee, who is conducting a study sponsored by RosettaNet and two of its board members, Intel and Arrow Electronics, Inc., of RosettaNet's potential payoffs to businesses. To put it more simply, says Hal Varian, dean of the University of California at Berkeley's School of Information Management and Systems: "If the electronics industry and RosettaNet can't do it, then few other industries will be able to follow." And that, he adds, would put a serious crimp in how fast—and how far—the Internet revolution can realistically be used by industries everywhere to wring out savings, boost productivity and integrate all industry players around a common goal.

RosettaNet backers are feeling the heat. The consortium's executives, who today represent some 400 companies from FedEx Corp. to Cisco Systems, acknowledge the problems, and amid the ongoing industry slump have been scrambling to turn things around. RN's Hamilton, who came on board in February 2000 to make changes, admits the organization was spinning its wheels for much of last year. She attributes the consortium's glacial progress to an overly ambitious development effort that has since been scaled back. Last year, she says, RN created a list of all the standards that would be needed to tie industry players to the Web and then began asking participants to use them. "But the list kept getting longer and longer," Hamilton says. "It was never going to end."

So what's RosettaNet's problem? First, early adopters have yet to agree on seemingly mundane matters such as standard ways to identify and describe common electronic components. Virtually every company has its own internally developed nomenclature for every item in its inventory. A simple capacitor, for example, is listed by one 14-digit inventory number at Solectron but by a completely different number at Hewlett-Packard. Nearly every company in Solectron's supply chain uses its own set of numbers for each item in its inventory—creating multiple ID numbers for the same product parts. "Automated trading won't go very far until companies can similarly describe what they're buying and selling," says UC-Berkeley's Varian.

Trying to take all of this global has been especially difficult. Two years ago, RosettaNet's members agreed to adopt a Global Trade Item Number to standardize product identification. But this spring the effort ground to a halt as component makers, distributors and contract manufacturers of computers quarreled over whose methods deserved to be standardized, among other issues. And until an accord can be reached, every company in the supply chain must maintain an ever-changing "look-up list" to match all the ID numbers of all the components carried by every company they do business with, slowing the process of doing business by hours, if not days, each time.

And there have been other roadblocks. Part of the RosettaNet challenge has been to develop standard definitions of individual electronic parts, but so far, three years after its launch, only some 3,600 out of a possible 10,000 or more "dictionary" definitions of, say, "price," "shipping," "processor speed" and "modem" have been devised.

RosettaNet also aims to standardize business processes across the supply chain, enabling one company's system to tap into another's system automatically online. Fifty-seven of a possible 100-plus business process standards, or PIPs, "are ready to go or are in late production," says RosettaNet's Hamilton. But most companies are using only one or two order management process standards in demonstration projects, and with only one or two partners out of a supply chain of thousands. Says Solectron's Jenkins: "This is a very big task, breathtaking in ambition. The amount of detail that is planned for the process and data definitions is very significant. Add to that the fact that more detail is going to be added to these process standards, and we're talking here about a very big effort that will take years and will never really be completed."

Companies need to change their existing networks to handle RosettaNet's process standards. "When your customer sends you an order and that order triggers the transaction, your system has to be able to handle that RosettaNet standard or definition, or it's not going to work for you," says Solectron Vice President and CIO Bud Mathaisal. But hardly anyone is prepared yet to let their existing networks, such as ERP systems, "talk" to each other automatically, without supervision—a goal of RosettaNet. At Solectron, for example, 95 percent of all electronic transactions are made using EDI, while only 5 percent use RosettaNet. And because there has been such little demand for RosettaNet, software vendors have been refusing to code RosettaNet standards into their applications, says Hamilton.

But the biggest problem underlying all of these roadblocks is culture—not technology, say Lee and other supply-chain experts. Shifting to RosettaNet standards, participants say, implies a loss of power and perceived control over their piece of the supply chain—and that's hard for many executives to accept. Companies involved in the RosettaNet experiment, Jenkins says, are still trying to "capture their own special processes and hide them as much as possible from the rest of the world. That's 180 degrees different from the RosettaNet strategy. There are significant internal acceptance barriers to overcome—and they will be overcome, but one executive at a time." In the Net-wired world, he says, knowledge is not just power, it is an asset to be appropriated, refined, enhanced and sold.

The lesson here for CIOs everywhere: Culture matters—and it must change to accommodate the cost-savings potential the Net can provide. "You can only go so far harnessing up your operation or industry to the Net before you have to start changing the entire organization and existing relationships among people," says Stanford's Mendelson. Adds Don Tapscott, co-author of Digital Capital: Harnessing the Power of Business Webs (Harvard Business Press, 2000), a guide to corporate survival in the new economy: "It's not easy for some of these guys to get naked. To succeed on the Internet, especially in these groups we call B-Webs like the virtual supply chain that RosettaNet is aiming to build, companies must prepare to reveal information about themselves, their products and their business processes, even previously sacrosanct information, and even to competitors, because it will be in their common interest."

But don't write off RosettaNet yet. Among RosettaNet CEO Hamilton's turn-around goals is a "milestones program" in which RN's largest original equipment manufacturing members—including Intel and Hewlett-Packard—have agreed to connect up to 100 mid-tier suppliers using at least one RosettaNet process standard. Hamilton is also planning to roll out other incentives to lead more companies to participate, including convincing software vendors to start writing and marketing RosettaNet-ready software to make it easier for companies large and small to standardize their business processes and parts designations. Many in the industry, including Jenkins, say this has been among the biggest reasons for RN's acceptance troubles.

Maybe so, but it's also true that it's impossible to wire up a paper-dependent industry overnight. RosettaNet is off to a slow start, and mistakes were made, UC-Berkeley's Varian says. But there's no turning back. "First the electronics industry," Varian says, "and next, the world." In other words, if CIOs are not looking to standardize supply-chain processes and language and wire them all up to the Web, they may have to be soon. Cost-cutting and adopting new technologies are the old way to gain advantage over rivals, but in the future, says business management guru Gary Hamel, it will be product and service innovation. "You can't be innovating stuff for customers if most of your brain matter and resources are wrapped around keeping inventories low and eliminating paper in your supply chain," he says. "That's old economy stuff."

All the more reason, say experts, for RosettaNet to get back on track. Says Mendelson: "If RosettaNet fails, it will not fail because the goal was flawed. It will fail because the culture lagged behind the potential of the technology."

Patrick Porter writes about technology from his home in Quincy, Mass. He recently completed a four-year stint as editor-in-chief of Software Magazine. Comments on this story can be sent to editors@cioinsight.com.



The RosettaNet Web Site

Achieving Supply Chain Excellence Through Technology

About.com listings and links

The Stanford Global Supply Chain Management Forum

Supply Chain Management Review



  • The more than 400 companies in RosettaNet have a combined revenue of more than $1 trillion.

Source: RosettaNet

This article was originally published on 07-01-2001