It's hard to place a value on knowledge management systems. Their ability to generate income is often measured indirectly; their links to cost savings frequently seem tenuous. The return on investment is hard to quantify. Too often, the case for implementing a system to leverage intellectual capital and expertise rests mainly on intuition: It seems like a good idea.
But intuition wasn't nearly enough to sell executives at Intec Engineering Partnership, a company whose dedication to thrift is exceeded only by its passion for sharing knowledge.
Intec is based in Houston and has offices in Argentina, Chile, Bolivia, England, the Netherlands, Malaysia and Australia. The privately held, US$80 million engineering and project management company serves the international oil and gas industry. Its 500-plus employees specialize in marine pipelines, terminals and facilities. Clients include BP, Exxon Mobil and ChevronTexaco.
CIO Fran Steele says Intec's culture is "extremely collaborative," having grown from four partners in one room sharing knowledge using 3-by-5 cards indexed by keywords. "It's a culture of sharing information, with a strong bias against bureaucracy and anything that might constrain creativity," she says.
During 2002, as Intec grew through expansion and international acquisitions, it was becoming more difficult to keep track of and access information. Intec wasn't alone. According to KPMG International, six out of 10 employees say difficulty in accessing undocumented knowledge -- such as the know-how people carry around in their heads and information that may be in documents scattered across a company -- is a major problem.
Early last year, a group of Intec engineers, later dubbed the Learning Team, volunteered to work on the problem of how to better capture lessons learned and share knowledge among Intec engineers. They diagramed how they solved engineering problems and envisioned an ideal process: An engineer with a question would go to a knowledge database that would either provide an answer or refer him to an expert. All new knowledge would be automatically captured and stored in the database.
The Learning Team decided that any technology to assist this process had to integrate existing knowledge resources, automatically find experts, capture results for reuse, facilitate the identification of best practices and provide a quick and easy user interface. Administration had to be minimal. "I wanted something that would not cost any more time of the people asking questions and answering," says Willem Timmermans, president and CEO of Intec. And ROI had to be clear from the start.
The Learning Team shopped around and selected software from AskMe in Washington as the product most likely to facilitate Intec's problem-solving model. AskMe agreed to customize a three-month, large-scale pilot of the system for 250 engineers.
The actual cost of Intec's system is proprietary, but Dan Wright, vice president of field operations at AskMe, says the average pilot includes 750 users at a cost of more than $100,000. The application starts at $375 per seat, with volume discounts beginning at 500 seats. There is also an 18 percent annual maintenance fee.
AskMe integrated its Web-based software and search engine with information repositories including 75,000 technical documents, Intec's existing skills and certification databases, and Lightweight Directory Access Protocol files of individuals' names, titles, locations, e-mail addresses and photos.
The Learning Team and AskMe developed a set of metrics that would be used to determine whether Intec would purchase the AskMe software after the pilot. The metrics included amount of activity on the system (visits, unique visitors, expert profiles viewed, questions asked, unique questions asked, answers given and average response time, for example), kinds of knowledge captured (such as questions answered, technical documents and best practices), system performance and, most important, user feedback and ROI.
The team agreed that qualitatively, knowledge management is about sharing knowledge, but quantitatively, it's about saving time. So ROI would be about time saved and putting a value on it (see box above).
The pilot, called AskIntec, began in May 2002. Three months later, it had exceeded all the performance and user metrics, and ROI calculations projected an annual return of 133 percent.
After nearly a year, the system is paying off almost exactly as projected. "Our numbers were pretty spot-on, but they're going up," Steele says, noting that the company estimates payback of 50 percent more next year as nonengineering employees are added and the system becomes embedded in the culture.
Actual savings are higher than the ROI figures indicate, she adds. An answered question, for example, often turns out to be worth much more than the Learning Team's estimate of saving 30 minutes, as senior project engineer David Raby demonstrated while working in Perth, Australia. Raby had an esoteric question about deep-water pipelines. Before AskIntec, getting an answer would have required accessing the library in Houston, ordering materials and having them sent through the mail. "I might have got the wrong stuff or needed additional stuff, and it could go back and forth for weeks," he says. "And I still wouldn't necessarily have the information I was looking for." Using AskIntec, he got 10 answers in a day, saving about three weeks' worth of effort, he says.
David Myers, a charter member of the Learning Team who is now Intec's knowledge-sharing manager, is working to sharpen the system's ROI metrics. A current upgrade will include an ROI engine that will ask users to estimate time saved on the spot rather than using formulas.
There are those who question measuring time saved as an approach to ROI. "If I could save every worker 15 minutes, I could say I'm saving millions, but does that time really go back to something productive?" asks Mike Gotta, an analyst at Meta Group. It's better to measure improvement in outcomes such as a decrease in defects or a quicker turnaround, he says.
More Than Numbers
Process metrics are difficult to implement at Intec because each engagement is unique, but the company is looking at process results anecdotally. "Some of the return on information is not quantified just by how quickly you can do something, but by the fact that you can do it at all," Steele says.
For example, during an engagement last year in Beijing, senior project manager Julio Daneri found that Chinese pipeline codes specified certain design parameters that his client didn't want to use. Daneri used AskIntec to query colleagues on cases where companies had successfully circumvented national specifications. Quick replies from engineers on three continents enabled him to build a case for using different specifications, without which his client wouldn't have been able to compete for the project.
The system is also improving Intec's sales process. Timmermans recounts that a prospective customer in Australia was skeptical about the local office's ability to draw on Intec's expertise all over the world. Intec invited him to pose a difficult question, which the Intec engineer put into the system. "The next morning they had four very relevant answers," Timmermans recalls. "We dazzled the client."
In the end, customers profit from Intec's knowledge management investment, Steele says, explaining that a typical oil facility can produce millions of dollars per day in revenue. "If we can cut weeks off a project and help them get their facility ready earlier, they can get to market sooner and get that revenue earlier," she says. "That's the ultimate value."
Time is Money
Intec's approach to calculating ROI began with defining and placing a value on the likely products of the knowledge system. The Learning Team defined four AskIntec products and assigned them values in terms of the engineer hours each would save. The value the team assigned to an engineer's time was a very conservative $50 per hour.
Intec estimated that these knowledge "products," when added to the knowledge-sharing database and then used repeatedly, would save engineers the following amounts of time:
-- Questions answered: One half-hour, or $25
--FAQs with answers: Five hours, or $250
-- Technical documents: 10 hours, or $500
-- Best practices: 50 hours, or $2,500 the results:
After the pilot, the Learning Team did the math. (Actual results are confidential. These figures are illustrative.)
--150 questions answered x $25 = $3,750
-- 25 FAQs x $250 = $6,250
-- 40 technical documents x $500 = $20,000
-- 4 best practices x $2,500 = $10,000
Total pilot ROI: $40,000
To estimate annual ROI, the team multiplied the three-month pilot's total value by five rather than four. That was to compensate for the gradual growth in use of the system and to account for the fact that only engineers were included in the pilot. (The system will be expanded to all employees.)
Annual ROI = $40,000 x 5, or $200,000