Drill far enough down into any supply chain problem and you almost certainly will find inconsistent and inaccurate data as a culprit. With processes that span multiple internal departments and external trading partners, supply chain management requires information exchanges between a plethora of parties and systems. Unless this information is all perfectly synchronized-and it seldom is-there are inevitable errors and inefficiencies.

While data integrity is an enterprise problem-and some companies are attacking it on an enterprise level with Master Data Management initiatives-the primary supply chain challenge is around product information. Having the same product named and described differently in different systems causes friction both in supply chain transactions and in partner relationships.

"The impact of 'out-of-sync' item information is felt everywhere that information is used in your and your trading partners' business," says John Stelzer, director of industry development at Sterling Commerce, Dublin, Ohio. "Whenever there is a discrepancy between the information your company has on file and the related information held by your customer or supplier, an exception occurs."

These exceptions can be purely informational, as in an incorrect invoice, or physical, as when an incorrect product is shipped, he explains. "Regardless of the category, each exception requires manual intervention and creates delays which increase the cost of doing business and undermine the quality of the trading partner relationship, and that directly undermines an organization's ability to effectively compete," says Stelzer.

Data inconsistencies also contribute to merchandising inefficiency and customer satisfaction issues, says Neeraj Gokhale, who founded Velosel Software, which last year was acquired by TIBCO, Palo Alto, Calif. "Companies try to place their products in retail stores according to demographic profiles," says Gokhale. "If they don't have accurate information about their products, they can't do the best job of merchandising them. And we all have had the negative customer experience of taking something to checkout and having it ring up wrong." The Velosel solution now is being marketed by TIBCO as Collaborative Product Information Management.

Experts say that these discrepancies also impair the adoption of collaborative supply chain processes and jeopardize emerging technologies like RFID. "With RFID, the idea is to identify the path of products along the supply chain, but you can never be sure you are talking about the same product if you haven't synchronized your data," says Steve Horton, chief marketing officer at RedTail Solutions, Westborough, Mass. "After investing a ton of money in RFID readers and technology, companies certainly need to make sure that the product they are talking about is the same one their trading partner is talking about."

Global Standards
The retail and consumer packaged goods industries have been working on this issue for years and are making solid progress through the Global Data Synchronization Network, which operates under the auspices of GS1 (formerly the Uniform Code Council and EAN). GS1 sets the standards that are used by companies to synchronize data, certifies data pools that actually send and receive data for member companies and manages the Global Registry, which is an automated directory service that identifies the data pool in which specific product information can be found.

"We have seen overall great progress and steady growth particularly in the last year," says David Garcia, vice president of marketing at 1Sync, Lawrenceville, N.J. 1Sync, a subsidiary company of GS1, was formed through a merger last August of Transora and UCCnet and is the nation's largest data pool, claiming to handle 95 percent of all synchronization activity. In total, there are 25 certified data pools around the world, including the Worldwide Retail Exchange and data pools operated by Sterling Commerce; Click Commerce, Chicago; and GXS, Gaithersburg, Md. GXS also offers a solution to manage data pools that has been adopted by four of the leading GS1 member organizations and is being used by data pools in 10 countries outside the U.S.

As of February, 1Synch had 455,000 registered items in its pool, a number that increases daily and is up from 168,000 a year ago, Garcia says. The number of companies exchanging data through 1Synch was 3,630 in February compared with a scant 176 at the beginning of 2005. "Data synchronization started with a lot of the larger multinationals but now we are really seeing a broad base of suppliers and retailers synchronizing data in the GDSN," says Garcia. In addition, the movement is moving beyond grocery where it has its roots. "We are seeing adoption begin in other sectors, particularly hard lines and office suppliers," he says. "Lowes, for example, has gone into production with more than 100 1Sync suppliers this year and we expect two of the largest office supply retailers to go live this year as well."

Grocery has a head start since it has been doing this since 1998, he says, "but the good news is that that there has been a tremendous amount of learning from their efforts and we will be able to onboard these other sectors much more rapidly."

RosettaNet Helps Synchronize Electronic Product Data

RosettaNet, another subsidiary of GS1, enables product data clarity among high-tech manufacturers and trading partners, though in a different way than the Global Data Synchronization Network. RosettaNet and its member companies develop e-business standards that align processes among trading partners and provide a common language for e-business transactions. "RosettaNet and GDSN focus on different types of data for different purposes," says Paul Tearnen, vice president of standards. "Both exchange product data, but while GDSN is concerned with attributes to help a retailer stock and sell the product, RosettaNet attributes are more focused on exchanges of information with suppliers that help engineers design the product." There is some overlap with GDSN, however, because in the electronics supply chain distributors and retailers need component information to sell the product. "Retailers like Best Buy need to be able to fill out those little tags that fit on the shelf that let consumers compare product characteristics and understand why this computer is worth $200 more than that one," Tearnen says. Before companies were able to easily exchange this information with RosettaNet, new products often would arrive at a warehouse before the product information, which meant that they could not be moved to the retail floor, Tearnen explains. Electronic products often are very complex, making it difficult to get consensus around what attributes should be included in the product description, Tearnen says. "RosettaNet pulls together a number of product information experts for that particular product category and they sit down and decide which attributes are important," he says. "This is challenging because they have to balance purchasing and marketing issues. The purchasing person wants commoditization and the marketing side wants differentiation." Often the most heated discussions are over new attributes that aren't generally recognized by the community as relevant. "The community might decide that a new attribute information has only marketing applicability and should not be included in a descriptor of product form, fit and function." RosettaNet has worked hard to accelerate this whole process so that new attributes can be added quickly to the standards "because the products change quickly," Tearnen says. A good example is a WiFi antenna on laptops, he says. That is a common attribute today, but it did not exist two years ago. The more than 500 companies that participate in RosettaNet report that its standards and implementation infrastructure have driven reductions in cycle times, reductions in inventory cost and improved productivity through automation.

Grocery retailer Wegmans is perhaps the most advanced user, having completed data synchronization with 416 suppliers that represent 82 percent of its cost volume. In addition, Wegmans is synchronizing at least some data with 956 suppliers out of 1,000 that signed up to participate, according to a company news release. Targeted departments for data sync include grocery, dairy/frozen, health and beauty care, cosmetics and basic general merchandise. "Focusing on key departments has helped Wegmans and our suppliers synchronize data effectively, and we believe the same approach will enable success across the industry," says Danny Wegman, CEO and a long-time advocate of data synchronization.

In March, five other major grocery retailers joined Wegmans in asking their suppliers to begin rolling out data synchronization on a category-by-category basis. The five are Associated Food Stores, Associated Wholesale Grocers, SUPERVALU, Unified Western Grocers and the all important Wal-Mart.

Moreover, retailers appear determined to make suppliers toe the line on the accuracy of their data. In a compilation of data synchronization case studies last year, Wegmans cited data accuracy as a continuing challenge, with a then-recent audit showing that only 35 percent of the item data it had received was 100 percent accurate. "The data that is sent for synchronization at Wegmans is used within all operation units," says Brad Papietro, e-commerce manager. "As a result, high quality data is a requirement to drive any benefit from GDS."

Prasad Vuyyuru, global solutions manager for retail and CPG at Infosys, Atlanta, underscores this point. "Say a manufacturer sends product size information that is off by only one-tenth of an inch," he says. "It doesn't seem very important, but when all the products are stacked next to each other, each small difference in dimension becomes significant."

Some retailers have instituted audit programs to provide suppliers with guidance and feedback on the quality of their data. Another way that retailers are forcing suppliers to adhere to accuracy requirements is by checking daily business documents like advance shipping notices and invoices to see if the information in those documents is consistent with the information the company published to them through data synchronization, says Stelzer.

"Manufacturers that choose not to first do internal synchronization, but just to publish their data to the retailers to satisfy them, will be easily detected as a retailer begins to check," he says. "So there is a push for those who tried to shortcut this to go back and be serious about data quality, which should have been the case to begin with."
"Synchronizing data with partners actually is the easy part," says Vuyyuru. "Streamlining internal processes and synchronizing internal systems to get to one version of the truth is the challenge."

Internal Synchronization
The difficulty in getting and maintaining one version of the truth for product data begins with the "decades of bad data stored in legacy systems," says Garcia. "A lot of these errors have become ingrained, so companies really have to roll up their sleeves and look under the hood." The good news, he adds, is that while this preparatory work might affect the pace of adoption, "it ensures better accuracy in the long run."

One facet of this problem is that the information that needs to be exchanged about products goes far beyond simple identification. It also must include numerous attributes needed by retailers to store, market and sell the product.

This includes weight and dimension, but also information like color, style, size and functional features. Increasingly, companies also are looking to include price and promotional information in published data.

All of those pieces of information are stored in various systems throughout an enterprise and sometimes in the systems of contract manufacturers or other partners as well.

"A company has to take inventory of its data assets and find all the places that data is being stored and operated on," says John Williams, vice president of services at Collaborative Consulting, Woburn, Mass. "Then they probably need to step back and divide those into tier-one applications, which actually have control of the data, and tier-two applications that only consume the data. Focusing on the tier-one applications makes the project more manageable." In a recent engagement, he says, a client started off with a list of 400 applications but was able to narrow that down to "around 30 that it really had to work on."

The goal is to create one "golden" master database as well as rules and processes for maintaining the currency and integrity of that database. This often raises data ownership concerns, says Kevin Prendeville, senior manager at Accenture, Reston, Va. "For this kind of initiative, it is imperative to get business unit buy-in as well as strong organizational support," he says.

"What many leading edge companies are doing is creating a data governance board," says Aditya Srivastava, director of MDM solutions at i2 Technologies, Dallas. "They have a cross-functional team that clarifies which group owns which information and they define policies and create software models around how that data is accessed and who can add, change or delete information."

At Procter & Gamble, an office of global data management is responsible for data stewardship. "We have a single global organization that manages the data and all the standards that go along with it," says Greg White, associate director of that office.

Data sync efforts at P&G started seven years ago with the first step of centralizing product master data into a single repository to get a global view of products, White says. "In so doing, we had to align to common data standards and common data management practices across all our countries and all our businesses. That was a big, big job-and it doesn't even speak to the issues of collecting data in a consistent manner and being able to use it in other systems as well as with our trading partners. We are still working on that."

P&G uses 1Synch to register its products and publish product information to retailers. "It has been a huge advantage to have an industry standard so we didn't have to invent this all ourselves," White says.

These efforts also proved very helpful in integrating P&G's acquisition of Gillette, says Milan Turk, managing director of global customer e-collaboration. "The fact that both P&G and Gillette were working independently to enable the sharing of synchronized, standards-based information with their trading partners made this a lot easier," he says.

Technology Support
P&G developed its own software to create and validate a central product data repository, but today many companies are turning to Product Information Management applications to help with this task. There are many players in this space, including major ERP providers and all of the vendors interviewed for this article.

	"Not a single one of our customers is interested in just one small portion of their data." - Cliff Longman of Kalido

Hardeep Gulati, senior director of product strategy for PIM at Oracle, describes the core steps of PIM functionality: (1) link to third-party and legacy applications and import data into a centralized data model; (2) transform the data by identifying and resolving duplicates and errors, then enrich the data with external information; and (3) synchronize the data by capturing all updates into the central master and pushing updates to all linked applications. While this is the most common approach, some companies prefer to shut down the databases in local applications and have all of them access the central database so that data entries are made in only one place. "One of the things we did early on at P&G was to strip out the master data from all of our transaction and reporting systems," says White. "So whenever the application needs master data it calls for it from our central database and gets it. That way we make sure it stays in synch."

PIM solutions also typically have a tool for creating business reports and other documents, which can include publishing item information to a data pool.

It also is important that a solution be able to scale, says Srivastava. He notes that i2's Product Information Management solution is part of its Master Data Management platform, which has a "very flexible data model" and is already installed at more than 100 of the largest corporations in a range of vertical industries. It has workflows "for managing the life cycle of the product, such as introducing a new product, changing and/or correcting an existing product as well as discontinuing a product temporarily or permanently from the supply chain," he says.

The importance of scale is reiterated by Johan Sauer, general manager of the master data management group at Click Commerce, Chicago. "Once you have content that is clean and accurate, you have to be able to store it in a way that is accessible, searchable and scaleable," he says. "And we are talking about serious numbers. Retailers really need to have capacity for millions of items in different categories." The Click Commerce technology, he says, enables product catalogs with 7.5 million items in multiple languages combined across different systems.

TIBCO's Gokhale emphasizes the importance of automating updates "as a matter of process, not as a matter of exception." TIBCO's Collaborative Product Information Management solution "has an eco-system that ensures that all changes are recorded in the master data and that the updated information is published to every one using that attribute, without anyone having to remember to take this action," he says.

One of RedTail's solution strengths is that it has built an adjunct application using SAP Workbench tools that sits within SAP," says Horton. "This offers SAP customers the ability to extend their SAP R3 to support the management of all product attributes needed for global data synchronization," he says. SAP alone doesn't have the ability to retain "more than a handful" of attributes, he says, but with the RedTail add-on, "companies have everything required to send GDS data in their SAP environment."

Sterling Commerce has a complete PIM solution, but Stelzer says he believes this is "overkill" for many companies seeking to get an accurate product data master and to synchronize that information with partners. As a result, Sterling last year introduced Commerce Item Synchronization. "This solution has all of the capabilities to handle the challenges of data synchronization, but it comes with pre-configured processes, so companies can't get started quickly and easily." The solution is built to enable companies to move up to a more complete PIM implementation as they become ready for that, he says. Sterling also offers companies a hosted data sync solution, he says.

Instill, Redwood City, Calif., is another company "providing a bridge" for companies and industries not yet ready for full data synchronization, says Jeff Smith, vice president of marketing. Specializing in the food service industry, Instill maintains an industry-wide cross reference of all the different ways that distributors refer to the same product. It maps against this the product lists of customers like Nathan's Famous. "Trying to map all of that manually was very difficult," says Nancy Murphy, vice president of purchasing at Nathan's. "Now we give Instill our data feeds and they map all the data and give us back a file that consolidates all those products."

"We support data pooling but the food service industry is not there yet," says Smith. "So we provide our customers the same business benefit in the near term."

Some companies get to PIM through the larger umbrella of Master Data Management. Kalido, Burlington, Mass., was developed as an in-house MDM solution for Shell Oil and then spun off as a separate company. "Not a single one of our customers is interested in getting a single version of the truth for just one small portion of their data," says CTO Cliff Longman. "They all have at least the ambition to implement a wider solution that will give them a single, clear version of customer data, product data, organization data, whatever, in order to drive operational systems in a more efficient and effective way."

Benefits Are Clear
Whatever the approach, and however difficult the journey, the benefits of data synchronization are clear and have been well documented in a number of case studies researched by Capgemini and sponsored by the Global Commerce Initiative. In summary, those benefits are: increased sales due to faster time to market and better on-shelf availability; improved productivity in the maintenance and publication of product masters and in fewer errors to reconcile; lower costs for transportation and deductions in the case of manufacturers and for transportation, store operations and planogram management in the case of retailers. (The entire report, "Global Data Synchronization at Work in the Real World," can be downloaded at www.gci-net.org.)

Milan says that P&G has been able to measure and capture data synchronization benefits "in terms of speed to market and of moving our products through the retail listing process much faster, thus getting them on the shelf faster and thereby increasing sales. We also have been able to measure and capture benefits from more accurate purchase orders that come from our customers. Because the customers have correct information about the product, we can enter it into our order management systems automatically and electronically. Those are today among the biggest sources of benefit, though realizing benefits is an ongoing process."

Gokhale also cites faster introduction of new products as a key benefit. One TIBCO customer that formerly took 12 weeks to collect, validate and get approval for the necessary data on a new product can now complete that process in two to three weeks, giving them nine extra weeks of selling time. "Using our technology, they have implemented a process that ties all of these steps together in a consistent fashion," he says. "At the end of the process, once the complete record has come together, the company uses data synchronization to send it to retailers, but at the same time they internally align the same data. So not only have they shortened the process, which is amazingly valuable, but they have become a lot more predictable because they have much better control."

Some unexpected benefits also are being realized, says Garcia. When the alcohol beverage industry added a product attribute identifying the country of bottling, for example, it became much easier for wholesalers to identify their eligibility for tax credits, a savings worth millions of dollars industry wide.

"This is just one of many unexpected side benefits we are seeing from data sync," says Garcia.