Rochester, NY--Lakes--a new family of products that can copy, print, fax, and scan--has transformed Xerox from a copier company into a leader in digital document technology.

"In effect, the mission of Lakes was to re-invent copying," says John Elter, vice president and chief engineer, new business development at Xerox's Office Document Products Group. Xerox sunk $450 million over seven years in the project--it was the company's big chance to join the digital information age and a "make-or-break-your-company project," says Elter.

"My contribution was to latch onto the idea, champion it, define it, and create a vision that the corporation could rally around and be willing to spend hundreds of millions of dollars on," he says. In the process, Elter changed the way Xerox does engineering. Lakes carried the banner in leading the corporation in the extensive use of quality-improvement techniques such as benchmarking, QFD (quality function deployment), and robust design.

The first Lakes product--the flagship Document Centre 265 Digital Copier (265DC)--debuted in September 1997 and its revenues have exceeded company projections. The Webster, NY-based factory makes about 250 units a day, and is adding a second shift to keep up with demand. Since the second quarter of 1998, revenues for Xerox digital products have outpaced those of analog, or light-lens, copiers. And according to market research firm International Data Corp. (IDC; Framingham, MA), the 265DC has made Xerox the number one supplier of digital black-and-white copiers.

This is just the latest fruit of Xerox's quality efforts. After losing major market share to Japanese companies in the late 1970s and early 1980s, David Kearns, former Xerox CEO, launched multiple campaigns to improve quality at every level of the company. In the early 1980s, Xerox began a program called Leadership Through Quality, and became the first American company to regain market share from the Japanese.

The programs paid off. Xerox is the only company to have won all three major quality awards around the world:

The Deming Award in Japan (1980)

But the company's business strategy hasn't always been so successful. Almost every book you read about Silicon Valley has an early section about Xerox PARC, or Palo Alto Research Center. PARC researchers invented the computer mouse, the icon-based computer interface, fax, high-speed computer networks, and the laser printer; but you'd never know it because other companies took these technologies to market and made successes of them.

More recently, Xerox was seeing its traditional copier market shrink. Fifteen years ago, 90% of paper printed in offices was processed through copiers. In 1997, almost half went through printers. The ongoing switch from analog technology to digital networked technology has been blurring the line between copiers and printers. Because digital machines can be networked, digital copiers could also work as printers.

"Given all that as a backdrop, the time had come for the copier to join the network and for Xerox to participate in the whole process--right through to commercialization," says Elter. At the dawn of the 1990s, he began rounding up engineers and support staff to create Lakes--a digital-copier platform to be designed from the ground up.

All the subsystems were new and unique. The architecture reduced the number of spare parts by an order of magnitude by using a limited number of high-level assemblies. Having these subassemblies self-align eliminated nearly all mechanical adjustments, reducing the time it takes to manufacture and service the product. Customer-replaceable units (CRUs), a subset of these assemblies, group together components that tend to wear out faster than the rest of the machine. Users replace the units on a regular basis, reducing downtime and improving customer satisfaction.

Consider the following additional features:

Scans and prints at 65 pages/minute, with high image quality.

The project was daunting, but Elter got behind it and saw it through to the end. "John is a visionary, a leader," says Ray Povio, chief technology officer for the Office Documents Products Group and the mechanical design leader for the 265DC. "John was like a gang of one. He sold management on investing in the project."

Xerox Chairman and CEO Paul A. Allaire was a part of the process almost from the very beginning. "Very rarely does a CEO of a big company come down and visit an engineering team regularly," says Elter. "And Paul would come at least once a year."

Allaire says the company's investment paid off. "The Lakes platform has been a critical factor in our successful transition to digital products," he says. "It reflects John's commitment to quality and reliability, his vision of creating a product that embodies customer requirements, and his leadership in converting that vision into an exceptionally well received product."

Lakes team members are even more enthusiastic than the CEO. For seven years, they--including Elter--worked long days, weekends, and holidays without any overtime pay. They were committed to the project and to delivering top-notch quality.

In fact, Xerox commissioned New Mexico-based corporate training company LivingSystems to write a book about the personal journeys or vision quests many employees undertook. These employees learned what quality meant and how to implement it, developed new ways of working together, and adopted new mindsets about the importance of setting and reaching both professional and personal goals. They changed Xerox corporate culture from the inside out, and set the standard for future projects.

To John Elter quality means fully satisfying the "voice of the customer"--what actual customers wanted from a digital copier.

The Lakes group spent six months listening as the first part of the design process. They held focus groups, did telephone interviews, and mailed surveys. They visited 20 cities and grilled hundreds of employees at all levels of customer companies about how they use copiers and what features their ideal copier would have. One surprise--customers wanted a copier on wheels to make it easier to move around. This feature--which Xerox incorporated--also makes servicing the 265DC easier for Xerox technicians.

"In the past, Xerox had been more of a push as opposed to pull company," explains Povio. "We wanted to have a pull system. There were several unsuccessful initiatives where we pushed products to the customer and found out in the end that that wasn't really what the customer wanted."

The voice of the customer was loud and clear. The three paramount wants were: 1. professional copy quality, 2. a machine that never jams, 3. a machine that gets fixed the first time. These three became top design goals.

These goals are called exemplars in a quality-planning process known as "quality function deployment," or QFD. It's the means of getting customer wants implemented in the final product. "QFD lets you take the voice of the customer and create specs, then critical parameters," says Povio. Xerox used QFD to deploy quality all the way to the factory floor. The process was so critical that each Lakes member took a three-day QFD seminar from the American Supplier Institute.

Benchmarking. Engineers were deprived of doing actual design work for six months by participating in a process called "benchmarking." The goal: to have the 265DC be the best in class in every possible category.

"We visited companies around the globe recognized as leaders in specific areas of quality and customer satisfaction," says Elter. "When we found the absolute best competitive system or service, we made that our benchmark--the standard we would meet or beat."

Povio was in charge of a formal benchmarking process. There were 27 benchmarking teams comprising engineering, manufacturing, and service people. Each was in charge of a major function or module of the 265DC and took a one-day seminar on benchmarking.

"Then we sent them to the competitive evaluation lab," says Povio. "Each 8- to 10-person team spent three weeks tearing down and testing all kinds of copier equipment. They looked at part count, ran tests, and got all this data--it was like the Starr report." The next step was to determine the best of breed for each function or module: Xerox, the competition, or a combination of both.

"At that point, it's just a concept," continues Povio. "Then we took them through value analysis/value engineering--a process for creating a concept you can design that meets all or most of your requirements." The teams did three iterations to ensure they had come up with the best value for the customer at the lowest cost and complexity and best reliability.

One way the Lakes team delivered low complexity and high reliability was by designing the customer-replaceable modules, or CRUs.

The Lakes team designed the CRUs to be easy to install and programmed the 265DC to prompt users to replace the modules on a regular basis.

This form of preventative maintenance prevents much of the downtime and customer dissatisfaction because the customer is in control of restoring the machine to its proper operation. The lifetime of a given CRU is shorter than the expected lifetimes of its parts.

Say a CRU has 150 failure modes. "Before, if any one of the module parts failed, it would require a service call," says Faull. "The service rep would come, take it apart, change that part, and put it back together." But there would still be 149 failure modes the rep hadn't addressed. Replacing the entire CRU sets the clock back on all those modes, notes Faull.

Robust design. Another key way the Lakes team implemented quality was through using "robust design"--a process developed by Japan's Genichi Taguchi for efficiently designing parts and products that are unlikely to fail. Robust design involves focusing on getting the desired output from a given input taking into account "noise"--conditions and circumstances the engineer cannot control.

"Ralph Faull was my alter ego," says Elter. "He instilled a major change in the way we do design. Not the mechanical design--i.e. choice of materials and so on--but change in the way we get the numbers for tolerances and dimensions that you see on engineering documents."

"John brought long-term commitment to change and was a key driver on changing the engineering work process," says Faull. "He gave me full authority to implement robust design," says Faull. Between April and June 1992, Taguchi himself trained 190 Xerox employees in robust design concepts.

"Traditional methods were that designers would conceive a mechanical design, make a drawing, put numbers on it, build it, test it, find out what didn't work, fix it, and go through another iteration. This resulted in very long development times," Faull says. "The concept here was to try to get the numbers before you build the product and really have confidence in those numbers."

In robust design, the build-test-iterate method was replaced by having the engineer consider what he or she wanted the product or module to do given its set of inputs and taking noise factors into account. "You formulate your ideal function into an input, output model," explains Faull.

"For example, a paper feeder's input might be the rotational torque of a motor; the output is the distance the paper is moving," explains Faull. "The noises would be all the different types of paper a customer could use--there could be thick paper, thin paper, transparencies, cardstock. And you'd want one rotation of that feeder roll to always move all those papers the same distance. These noises are customer usage conditions. Mother Nature kinds of noises are conditions like temperature and humidity."

Engineers would create a fixture, and then make all the adjustments so different types of paper would feed as uniformly as possible while the design is still in this crude form. And just by moving the voltage up and down or changing a gear, says Faull, you could make sure the feeder would work if the machine were redesigned to make 40 copies/minute or 60 or 90.

This tack saved lots of engineering time. "We had a strategy for 22 different products, all with different features," says Elter. "We had to take all of that into account for the initial design, so that the first design did not exclude the future." In fact, the first design called for the 265DC to work at 40 copies a minute, then it went to 50, then 60, then to 65. Faull credits robust design for being able to change speed with minimal design or manufacturing changes--a critical factor for a multiple-product platform.

"Using these methods, people have said the first iteration is like the third or fourth using traditional methods," says Faull. Lakes was the first product platform at Xerox to have such a large focus on robust design to reduce part variability for a wide range of products. It's now standard practice throughout the company.

The 265DC has about 250 replaceable parts. "The fuser is one part in terms of the assembly of the machine, but it probably has 80 to 90 parts in it," says Povio, the mechanical design leader for the project. The vendor is responsible for assembling a part, testing it, and certifying it before it comes into the factory. "This is how we can get the machine put together in an hour with a high level of quality and also keep the spare parts that a service rep carries around to a very few," he says.

A traditional factory would take twice the floor space, says Whitmyre. He keeps track of quality metrics and displays them in a common area. And people actually look at the numbers to see how the factory's doing.

"How do you manage such a large complex program while ensuring quality? That's the $64,000 question," says Elter, but he's got the six-point answer.

First, create a vision people can rally around

Elter is currently on sabbatical at MIT studying complexity theory, microelectromechanical systems (MEMS), and managing the emergence of new technologies. He lectures at the Sloan Business School on product design and how to manage large projects--which he says is sort of a subset of complexity theory. He should know.

REMOTE DIAGNOSTICS

The 265DC's Sixth Sense software lets Xerox service representatives analyze the machine remotely over an ordinary telephone connection. About 600 data points are analyzed for performance and wear. This information lets the rep arrive at the customer location with a good idea of what's wrong and the necessary spare parts, minimizing downtime. Sometimes a visit isn't even needed--a technician can make certain electronic adjustments to a customer's 265DC via computer over the phone line.