These days, performing a search on the thousands of documents typically stored in a large corporate database such as the one at Procter & Gamble Co. is a fairly routine task. The user plugs in key words and the search engine spits out a list of relevant files.

But a not-so-routine search function is the ability to find an expert within the company along with the desired documents. Procter & Gamble’s Innovation Net (INET) attempts to bridge that knowledge gap by linking searchers with experts within the company. INET reaches 18,000 key knowledge workers at Procter & Gamble in departments such as research and development, engineering, purchasing, consumer and market knowledge, patent/legal and business information systems.

When INET was established four years ago, it was designed to enable users to perform searches to find the information that was already documented — marketing data, research and development memos and other information typically created in the course of business. The system also enabled users to see similar questions that have been asked regarding a particular topic and the response from the experts.

Last year, P&G added software from AskMe (www.askme.com), enabling INET to evolve into a place where users can search for experts within the company along with documents. Following a successful pilot, the consumer goods giant integrated AskMe Enterprise software into the intranet’s framework to locate, catalog, transfer and maintain employee knowledge. “The quality of conversations going on was very high and could be tied to people moving projects forward. It was clear that this would be a pretty good investment,” according to Mike Telljohann, associate director at P&G’s technical center in Cincinnati.

With a company as large and as global as P&G, it is difficult for workers to easily tap into information about a project that might be relevant to their own project. Someone in engineering or marketing for another product may have tackled a similar problem six months ago, but they’ve never crossed paths with the worker who is currently dealing with the issue.

AskMe performs natural language processing to extract key concepts the user is looking for. The natural language processing identifies key nouns and pronouns in the user’s search. Fuzzy matching identifies the root of each word to present acronyms and synonyms.

The system applies intelligent pattern matching that accounts for where words are placed. For instance, the accuracy of the match depends on how close the words are to both each other and the subject of the document or answer.

Based on how much particular workers are involved in certain subjects, and the profiles they have filled out, the system forms a directory listing of individuals designated as subject-matter experts. Users are presented with an interface that looks similar to a personalized Yahoo! page. “Category champions” are charged with updating information in their specialization.

According to Telljohann, efficiencies have been gained by establishing a single knowledge base in the company. “Experts see great value in being able to see a question once and refer repeat questions to those answers,” Telljohann notes. In addition, the integration of the system with e-mail enables immediate exchanges.

Smarter Than Yourself
by Dennis Eskow

While few consumer goods companies would reject the opportunity to adopt a best practice, only the rarest of winners has gotten beyond best practices and into the heady area of knowledge management.

“The trouble with best practices is that they are based on past performance,” says Tom Galinda, VP of process engineering at Pantheon Foods, a British-based manufacturer of Mediterranean packaged foods. “They try to package knowledge. But smart companies need to manage knowledge. That means having a discipline that knows when knowledge is coming in, where it’s kept, how it’s pulled from one place to be used in another.”

P&G Gets Tools
The age of knowledge management is upon us, and CG companies are beginning to approach the subject cautiously. Procter & Gamble Co., for example, has acquired some KM tools to identify subject experts in various departments and make their knowledge available to others. P&G’s approach is through an employee education portal that allows project workers to collaborate in everything from manufacturing to research. At first, the portal contained best practices information. In a recent interview, P&G technical center associate director Mike Telljohann said, “People often didn’t know where to go with questions or issues. They suspected, given the size of the company, that there was more out there that they just didn’t get to leverage.” P&G Adopted AskMe Enterprise knowledge-sharing software, an Intranet tool that forms subject directories so that a researcher in a Cincinnati beauty products lab can communicate with a California electronics expert on the effects of sunlight on cells.  (For more on P&G’s KM system, see “Expert Search,” page 30.)

Knowledge management, which uses a variety of specialized technology tools, is not a technology itself. One of the field’s foremost experts — Yogesh Malhotra, founder of the Virtual Library on Knowledge Management (www.brint.com) —urges companies to develop a knowledge program before considering technology. But a properly managed program, he says, helps you make your own knowledge obsolete “before others obsolete it.”

Smart organizations can use existing technologies to manage knowledge. DemandTec, for instance, has helped several consumer goods companies reevaluate their approach to KM through the adoption of Demand Chain Optimization technology. The best description of how it manages knowledge comes from a retailer, Nick D’Agostino of D’Agostino Supermarkets. “You have to let the software tell you things based on what it learns,” he says. With Demand Chain software, he learned that some of his stores should be stocked as convenience stores rather than groceries. This software gathers supply chain and customer data and with the help of mathematical algorithms, extrapolates actions that should be taken to improve the bottom line.

Campbell’s Knowledge Soup
Another instance of KM through non-KM software comes from Campbell’s Soup, which deployed PeopleSoft two years ago to speed up the exchange of knowledge about government regulations and the company’s compliance with them. Here is an area where global manufacturers have almost daily nightmares. Each human resources manager around the globe uses the system to monitor progress on regulatory compliance.

Campbell’s relies on the Web readiness of the PeopleSoft tools. And it is the Web that has made KM a more viable tool for corporate managers. Thanks to Web-related technologies, companies have learned how to identify and capture knowledge when and where it is needed.

Sorting the issues is a challenge, and several centers of KM have sprung up. The one with the best focus on manufacturing and consumer goods is the University of Texas (www.bus.utexas.edu/kman/). UT and Malhorta’s organization can help shed light on the plethora of specialized KM products. These range from wide-scope to narrow scope software suites.

A wide-scope solution from Primus (www.primus.com) allows corporations to store knowledge in one server, clean the data periodically, and provide customers, employees and suppliers with Web-based customer service type answers to knowledge searches. Knowledge-based Systems (www.kbsi.com/) analysis and modeling in a more narrow-cast approach to the collection and use of employee intelligence. Several companies have adapted their data warehouses into KM tools.