The Search for Quality

TQM, Capability Maturity Model, Six Sigma -- they're all about quality and the IT community's repeated attempts to get software development right. "Quality has been a focus since the beginning of time -- '65 or '70," says Howard Rubin, executive vice president of Meta Group Inc., a Stamford, Conn.-based IT consulting firm. "We've basically had 30 years of quality repackaging -- new and improved."

Yet Rubin's research shows that for all the quality efforts of the past 35 years, companies still spend more on software fixes than they do on development. "Running late and over cost is normal," he says, "and defects in software are considered state of the art."

Technology first caught the quality bug in the mid-'70s, when pioneers at IBM began to adapt the methodologies of quality guru W. Edwards Deming, who was working with U.S. auto manufacturers at the time. "Deming focused on building it right the first time," says Watts Humphrey, who was at IBM during that time and later became director of software quality and process there. "The fundamental quality problem is that most organizations build poor-quality products and try to fix it later."

In tech circles, Deming's work morphed into total quality management (TQM), though Deming distanced himself from the term, according to Joyce Orsini, who worked closely with him. "Most of what passes under the label of TQM is a very small part of what Deming had in mind, and in some cases is not related at all," says Orsini, now associate professor of management systems and director of the Deming Scholar Program at Fordham University in New York.

TQM sought to improve quality through ongoing refinements and continual feedback. "TQM tended to encourage continuous incremental improvement -- lots of small changes," says Tom McCarty, director of consulting services at Motorola University, the organization that drives learning, leadership and performance at Schaumburg, Ill.-based manufacturer Motorola Inc. Unfortunately, TQM often failed to live up to its billing because it was frequently delegated to lower-level quality managers who lacked the company overview and the authority to make significant improvements. But it lives on in the International Organization for Standards' ISO 9000 series on quality.

Capability maturity model

By the mid-'80s, another quality movement was developing at Carnegie Mellon University's Software Engineering Institute (SEI). Humphrey, who had become head of the process program at SEI, began developing the Capability Maturity Model. "The whole idea of CMM was that I had seen what good management, planned schedules and orderly ways of working could produce," he says. His work was based on a "Eureka!" moment at IBM. "One idea hit me like a ton of bricks: Nobody but software focuses on testing. The way people are trained, the way it's managed, is to bang out code and then fix it," he says.

Software professionals think quality is the responsibility of the testing department, he says. "But if you don't feel personally responsible for the quality of your work, you'll never get quality work."

Despite the SEI's efforts, Humphrey's insights didn't catch on as broadly as he'd hoped. "People didn't seem to get it and still don't," he acknowledges. "The software community really still believes testing is the only way to do it."

Six Sigma

In 1987, engineer Bill Smith upped the quality ante when he introduced Motorola to the concept of Six Sigma, which focuses on understanding customer requirements, aligning work processes to achieve them, then using analytical tools to improve those processes.

Six Sigma and TQM have a lot in common, but there are important differences. TQM seeks to incrementally improve processes. Six Sigma focuses on breakthrough improvements yielding bottom-line business results. TQM seeks improvement anywhere, but Six Sigma focuses on critical core processes to achieve maximum bang for the buck. That means Six Sigma has a big advantage because it speaks the language of business. "In Six Sigma, executives determine improvement opportunities, and they are more willing to sponsor projects, eliminate barriers and provide resources," McCarty says.

Regardless of the quality program, the companies that have had the most success in IT quality drives have tended to be those in manufacturing, such as Motorola and General Electric Co., says Rubin. That's because the manufacturing culture is all about metrics, while many IT shops still think of software development as an art. "Putting a quality movement in the world of software is like putting quality control on painting," he says. "It takes a change in culture and perspective."

Rubin says he thinks today's quality programs combine the best of what has come before. "They're more expansive; they're integrating more process, more measurement. They integrate better with the business environment," he says. That's essential, because quality efforts work only when people understand the bottom-line advantages, he adds.

McCarty says that bodes well for Six Sigma. "More and more, quality will be seen as a strategic tool for achieving business results," he says. And Humphrey agrees that true quality always serves the bottom line. "The best quality programs always save money and cut time," he says. "If they don't, they're not being done right."

Rubin says he hopes a new approach to hardware quality will spread into the software world as well. "It's called the 'No 9s' movement," he explains. "Instead of 'high 9s' availability [for example, 99.9999% uptime], they're talking about '10' availability [no downtime ever]. They're designing from the ground up so defects don't happen. It's a change in perspective."

Melymuka is a freelance writer in Duxbury, Mass. Contact her at kmelymuka@earthlink.net.

Copyright © 2002 IDG Communications, Inc.

7 inconvenient truths about the hybrid work trend
Shop Tech Products at Amazon