This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
1
ANALOG DEVICES, INC. (A)
Analog Devices Inc. (ADI) was a leading manufacturer of integrated circuits that convert
between analog and digital data. From 1981 through 1996, ADI experienced periods of growth
and stagnation, both achieving record profits and sales a experiencing its first loss ever. To meet
the needs of the changing market, management at ADI introduced a number of different
management tools to implement change. One such tool was its corporate scorecard.
ADI's corporate scorecard was recognized as a management best practice in a survey the Nolan-
Norton Group conducted in 1991. Despite this accolade, ADI's management was wondering in
1996 how to change the scorecard to best fit the needs of management, specifically, how fast to
change it and how best to use it to focus management attention in the future.
Analog Devices was founded in 1965 in Cambridge, Massachusetts, by Ray Stata and Matthew
Lorber. Stata had a B.S.E.E. and an M.S.E.E., both from MIT. In 1996 the company operated
predominantly in one industry segment: the design, manufacture and marketing of a broad line of
high-performance linear, mixed-signal, and digital integrated circuits ("ICs") that addressed a
wide range of real-world signal processing applications. The company's principal products were
divided among four classifications: general-purpose, standard-function linear, and mixed-signal
ICs ("SLICs"); special-purpose linear and mixed-signal ICs ("SPLICs"); digital signal processing
ICs ("DSP ICs"); and assembled products. SLICS were the largest product segment for the
company, accounting for 65 percent of total sales.
Nearly all the company products were components that typically were incorporated by original
equipment manufacturers (OEMs) in a wide range of equipment and systems for use in
communications, computer, industrial instrumentation, military/aerospace, and high-performance
consumer electronics applications. The company sold products worldwide; in 1996 one-half the
sales came from outside the U.S.
Analog Devices, Inc. (A)
2
INDUSTRY BACKGROUND
Real-world phenomena—temperature, pressure, sound, images, speed, acceleration, position, and
rotation angle—are inherently analog in nature, consisting of continuously varying information.
Analog sensors can detect and measure this information. The signals are usually converted to
digital form for input to a microprocessor, which is used to manipulate, store, or display the
information. In many cases, the signals are further processed using a technology called digital
signal processing. In addition, digital signals frequently are converted to analog form to provide
signals for analog display, sound, or control functions. Collectively, these manipulations and
transformations are known as real-world signal processing.
Significant advances in semiconductor technology have led to substantial increases in the
performance and functionality of ICs used for signal processing applications. These advances
include the ability to create very large scale integration (VLSI) mixed-signal ICs that contain
both high-performance analog circuitry and large amounts of high-density digital circuitry. The
analog circuitry portion of the IC is used to manipulate real-world signals while still in analog
form and to convert analog signals into digital form (or vice versa). The digital portion is used to
further process analog signals after they have been converted to digital form.
COMPANY STRATEGY
In 1996 Analog Devices' strategy was technological leadership. The company wanted to be first-
to-market with new products that had superior performance features. Analog Devices was one of
the world's largest suppliers of SLIC products.
During 1990–1996, Analog sought to balance its traditionally stable SLIC business with growth
opportunities for SPLICs and DSP ICs. The company built upon its expertise in linear IC
technology, developing special-purpose linear and mixed-signal ICs tailored to specific high-
volume applications in target markets. Analog also extended its expertise in analog signal
processing and data conversion to develop DSP ICs. These DSP ICs and its SPLICs addressed
the emerging demand for high-performance levels in many communications, computer, and other
high-volume applications. These products had a high level of functionality (i.e., many functions
on one chip) to satisfy OEMs' requirements for an integrated solution with low cost per function.
Analog Devices, Inc. (A)
3
To build upon its position as a leader in real-world signal processing, Analog was pursing the
following strategies in 1996: (1) Expand traditional SLIC business; (2) Become a major supplier
of general purpose DSP ICs; (3) Pursue growth opportunities for system-level signal processing
ICs; and (4) Leverage core technologies to develop innovative products.
TOTAL QUALITY AT ADI, 1983 TO 1986
From the inception of the company in 1965 till the early 1980s, sales at Analog Devices grew at
a rate of 27 percent a year. Yet in 1983, Ray Stata recognized that ADI was having problems
with the quality of its production. Its on-time delivery record was under 60 percent. Its process
yields, in some cases, were as low as 10 percent.1 ADI's customers were complaining about
quality, and its competitors had on-time delivery records and yields well above ADI's level. At
this time, Stata attended Philip Crosby's Quality School. This was ADI's first introduction to the
concept of total quality management (TQM).
Interested in implementing TQM but not wanting to add additional staff to create a quality
improvement function, Stata charged the human resources division with establishing a TQM
program at ADI. The first TQM effort never got beyond managers trying to become TQM gurus
on their own by reading books and going to seminars. As one general manager said, "I was
focused on growing the business, not on TQM." 2
By the end of 1984, ADI's sales had reached $313 million. During fiscal year 1984, revenue had
grown by 46 percent, profits by 105 percent, and orders booked promised another record year in
1985. ADI management felt it was in the middle of some of the fastest growing segments in the
economy. Many in the company were starting to talk about ADI becoming a $1 billion firm by
1988.
Unfortunately, between the end of 1984 and the end of 1986, sales had only grown 6.7 percent,
and profits had fallen by 38 percent. As Stata stated, “[F]or the first time, between 1982 and
1 Howell, Shank, Soucy, Fisher, Cost Management for Tomorrow, Financial Executives Research Foundation, 1992, p. 128. 2 Unless otherwise indicated, the statements in this case attributed to Analog Devices' Ray Stata (CEO and chairman), and Goodloe Suttler (vice president of strategic planning, quality improvement, and corporate marketing), were based on interviews.
Analog Devices, Inc. (A)
4
1987, we missed our five-year goals—and by a country mile. True enough, like other
semiconductor companies we were affected by the malaise in the U.S. electronics industry and
by the strong dollar. But the external environment was only part of the problem: something was
also wrong internally, and it had to be fixed."3
The factory was missing over 40 percent of its committed delivery dates. When 20 executives
with regular customer contacts were asked, "When the phone rings and it is an angry customer,
what did he say?" The executives responded, "The customer said, 'Where's my order?!'"
The defect level of product that reached the customer was more than 20,000 parts per million
(PPM). Competitors such as Motorola were achieving results under 1,000 PPM. Furthermore, the
poor quality caused a substantial amount of waste at ADI, such as front-to-back IC yields of less
than 15 percent, meaning that only 15 out of every 100 ICs that ADI started made it through the
process. These were well below industry yields.
Although in 1985 ADI's analog IC sales declined by about 5 percent, 1986 industry wide analog
IC sales grew by 25 percent. The analog circuits industry had returned to growth, yet ADI's
revenues were stagnant and its profitability was declining.
THE QUALITY SPECIALIST
In 1986 ADI hired Art Schneiderman as vice president of quality and productivity improvement.
Schneiderman had been a consultant with Bain & Co., where he had been directly involved in
establishing many quality improvement programs. He was seen as someone who could link ADI
to the "mainstream of experience and knowledge that is rapidly accumulating in [TQM]" and be
a teacher who could "help [ADI's] managers become more expert practitioners."4 Stata wanted
the quality improvement process (QIP), as ADI called its total quality program, to become a way
of life at ADI.
3 Ray Stata, "Organizational Learning—The Key to Management Innovation," Sloan Management Review, Spring 1989, p. 63. 4 Ibid., p. 63.
Analog Devices, Inc. (A)
5
Many of the general managers at ADI were skeptical of this new quality program, having
undergone the earlier, unsuccessful quality program. Additionally, they believed the quality
goals and the company's financial goals were in conflict. The financial basis of ADI's incentive
and performance measurement systems reinforced this belief.
HALF-LIFE
Schneiderman believed that "any defect level, subjected to legitimate QIP, decreases at a
constant rate, so that when plotted on semi-log paper against time, it falls on a straight line."5
The result is that every process can experience a 50 percent reduction in defects at a consistent
time interval. Schneiderman called this the half-life of the improvement process. As Sterman et
al. note, "The basis for the half-life dynamic is the interactive learning loop at the heart of TQM.
Improvement teams identify the root causes of defects, rank them in order of importance, then
propose, design, test, and implement solutions using the Plan-Do-Check-Act or 'PDCA' cycle.
The team continues to cycle around the learning loop until the root causes of most of the defects
are corrected, then moves to the next most important source of defects."6
Schneiderman had collected data on improvement activities. Exhibit 1 shows three examples of
the data he collected. The examples are shown on log-linear graphs to capture the full effect of
the half-life concept. By plotting improvements this way, it was easy for someone to see the line
indicating the improvement rate. As Exhibit 1 illustrates, each process had its own unique rate,
determined by finding the slope of the line fit to the data. This unique rate was the process's half-
life. (Appendix 1, The Half-Life/Complexity Matrix, and Appendix 2, Relationship of Half-Life
to the Experience Curve, give additional background on the half-life concept.)
Using the 1987 five-year plan as a tool, Schneiderman introduced goals for a series of quality
measures (Exhibit 2) that corresponded to what he considered to be ADI's critical success
factors: having innovative, high-quality products and being a reliable, responsive supplier. The
5 Arthur M. Schneiderman, “Setting Quality Goals,” Quality Progress, April 1988, p. 53. 6 John D. Sterman, Nelson P. Repenning, and Fred Kofman, "Unanticipated Side Effects of Successful Quality Programs: Exploring a Paradox of Organizational Improvement," Management Science 43, 4 (April 1997), pp. 504–505.
Analog Devices, Inc. (A)
6
goals were determined by combining customer demands with realistic expectations of each
measure's half-life.
He proposed reductions such as dropping process defect levels from 5,000 PPM in 1987 to fewer
than 10 PPM by 1992. Many managers just laughed at him. Stata recalled, "The first reaction of
our organization was to recoil from what looked like unrealistic objectives. But we reminded our
managers that if a company really gets its quality improvement act together, there is no
fundamental reason why these goals cannot be achieved."7
ADI'S SCORECARD
Schneiderman put together a single-page scorecard that showed three categories of
measures: financial, new products, and QIP (Exhibit 3). Measures within these categories
indicated how well ADI was moving toward its goals. The scorecard was prepared once a
quarter.
ADI considered the scorecard to be a breakthrough because it condensed pages of reports
into a simple, single report. It measured each critical success factor as well as financial
performance.
Schneiderman added to it the half-life and target for each of the measurements for the next few
periods. He did this to provide a link between short-term results and ADI's long-term plans, such
as improving on-time deliveries to 99.8 percent by 1992.
MAKING THE SCORECARD WORK
Schneiderman developed the following rules about how to construct the scorecard:
• The entire scorecard had to fit on one 8 ½” by 11" sheet of paper.
• The font size had to be 12 points or bigger.
• There were to be six times as many nonfinancial measures as financial measures.
7 Ray Stata, "Organizational Learning—The Key to Management Innovation," Sloan Management Review, Spring 1989, p. 70.
Analog Devices, Inc. (A)
7
In addition to the scorecard and TQM, Schneiderman helped change ADI from doing five-year
planning every five years to doing five-year planning every year.
He also helped create divisional scorecards. The measures for most divisions overlapped; all
could be tied directly to ADI's overall scorecard. The company allowed each division to use the
same scorecard or a unique one. By tailoring scorecards to each division, Schneiderman gave all
divisions the means to negotiate their goals and determine the appropriate half-lives of their
measures.
In 1988 Schneiderman began to roll out the balanced scorecard to the entire company. Each
division gradually developed its own scorecard and had successive levels develop theirs. They
also were not required to have unique scorecards. Lower-level management scorecards typically
placed less emphasis on financial measures than on nonfinancial measures on which managers
could have an impact. Each division was required to share the quarterly scorecard results with all
its employees.
Division results for each scorecard measure were compared. Exhibit 4 shows how different
divisions compare in on-time customer service performance results. Each division was shown
together, and the slope, or half-life, of the improvement was shown. In addition to distributing
reports like Exhibit 4, Schneiderman compared the scorecard performance results with the target
results during executive council meetings. He pointed out large favorable variations (which he
circled in green) and large unfavorable variations (circled in red), and asked divisional general
managers to explain the causes.
According to Goodloe Suttler, a graduate of the Tuck School at Dartmouth College and a vice
president and general manager of ADI's semiconductor division, the company used the corporate
scorecard as a communication tool. To the employees, it said,
• Measurement is the key to determining success.
• You cannot know how well you are doing unless you have measures.
• Here is what is happening in your division/plant.
To management, it said,
Analog Devices, Inc. (A)
8
• These scorecard items are the metrics of success.
• Focus on the items critical to success.
• You must meet objectives measured in the scorecard.
By 1991 the corporate scorecard was being used aggressively on a daily basis throughout the
organization.
PERFORMANCE8
Analog Devices showed dramatic improvements in its quality measures by 1990. Between July
1987 and July 1990, on-time delivery increased from 70 percent to 96 percent, cycle time
decreased from 15 weeks to 8 weeks, average yields improved from 26 percent to 51 percent,
and defects in products shipped declined from 500 PPM to 50 PPM. Other variables did not
improve during 1987–1991. Product development time did not decline. The stock price dropped
from $25 in 1987 to $9 by 1991, a much larger decline than performance by the market as a
whole or the semiconductor industry in particular. In 1990 Analog Devices experienced its first
loss (Exhibits 5 and 6). It missed its profitability goal for 1991 by 50 percent and suffered its
first-ever layoff of 600 employees (12 percent of its workforce). In response to the financial
crisis, Analog decided to shift its strategy of focusing predominantly on the SLIC business to
emphasizing the growth areas of SPLICs and DSP ICs.
CHANGING ROLES
During this time, ADI's management directly under Stata changed considerably. The company
promoted Jerry Fishman to president and named eight new vice presidents, including five from
outside ADI. A number of longtime vice presidents retired. The changes reflected ADI's efforts
to infuse a different culture. As one of the new vice presidents from the outside said, "Analog
had a product orientation, not a customer orientation. The financial dip helped bring dramatic
changes."
8 John D. Sterman, Nelson P. Repenning, and Fred Kofman, "Unanticipated Side Effects of Successful Quality Programs: Exploring a Paradox of Organizational Improvement," Management Science 43, 4 (April 1997), pp. 503–521.
Analog Devices, Inc. (A)
9
Following Stata's lead, the entire senior management team stepped up to the role of quality
leaders to demonstrate that improvement is everyone's responsibility. Schneiderman had been
ADI's torchbearer during the era of the quality specialist; now it was time to begin the era of
senior management. In mid-1992, Art Schneiderman resigned from ADI and passed the torch to
the new management team.
TQM OFF-TRACK
TQM's primary focus on cost reduction left many managers feeling most cost reduction was
done by the early 1990s. Many also believed the investment necessary to stay with, and
constantly improve, the TQM program outweighed the many advantages of continuing it.
Some at the company simply had no faith in the program. "There is some closeted cynicism
about quality [programs] in the company," noted Stata. "Among the engineers, it isn't even
closeted. They think it's crap."9 The end result was increased pressure to abandon TQM.
By the mid-1990s, ADI was bouncing back financially. Management believed that QIP had
improved the firm's profitability by reducing waste but, because it was primarily a cost reduction
tool, did not credit it with ADI's growth. In fact, although ADI was experiencing high growth
and profitability, many of the measures of quality were declining. The program appeared to be at
a standstill. Management still recognized TQM's value but felt that it was insufficient to address
the new problems facing the company.
CHANGING SYSTEMS
After Schneiderman left, many of the systems he had put in place changed or withered away. In
Goodloe Suttler's opinion, there was a "performance paradox." Borrowing ideas from Professor
Marshall Meyer at University of Pennsylvania's Wharton School of Management, Suttler
believed that all performance measures would eventually degrade. Performance would improve.
Variability would be reduced to the point where further improvements were of little value.
People would game the system. As a result, new performance measures would need to be
constantly introduced. According to Meyer, with new performance measures driving faster
Analog Devices, Inc. (A)
10
change, "accelerated learning rates suggest we will cycle through measures with greater
rapidity." Simply put, the better your measure helps you to improve, the sooner it will lose its
value. As Stata observed, "We are now recognizing as we get more sophisticated, it is harder to
get numbers that are meaningful."
ADI began to look for new tools that could make the numbers more meaningful, particularly
numbers that were leading indicators of value growth. As Suttler said, "The big problem with
TQM is that it has little to say about business strategy. TQM works well at stopping wealth-
reducing activities, but wealth creation doesn't naturally come from TQM."
Suttler pointed to evidence that TQM had been successful at ADI, such as the reduction of
outgoing electrical defectives from more than 20,000 PPM in 1987 to fewer than 50 PPM in
1994, and the improvement of front-to-back IC yields from fewer than 15 percent to more than
60 percent in the same period. On the other hand, he noted that ADI started TQM at a point
where the cost of waste was 25–35 percent of sales and reduced that to under 15 percent in seven
years. ADI believed that reducing the cost due to waste to 3 percent would take another seven
years. Although management considered cost reduction important, it was considered less critical
than finding ways to grow revenues. Using a model of dynamic complexity (see Exhibit 7),
ADI's management concluded that growing revenue was a more difficult process than reducing
cost and would take longer to implement.
HOSHIN
ADI began searching for new methodologies to create wealth. Ray Stata learned about a
technique called Hoshin kanri as part of his participation in the Center for Quality Management.
For Stata, Hoshin was an extension of the QIP effort at ADI and a realistic approach to center the
company's energies on wealth creation. Its main idea was to focus improvement on one or two
breakthrough objectives for the company.
Hoshin, as described by Suttler: "literally means policy deployment and control. Hoshin tells us
to focus on the most important objectives. Analog Devices has two: on-time delivery and new
9 Rahul Jacob, "TQM, More Than a Dying Fad?" Fortune, October 18, 1993, p. 66–72.
Analog Devices, Inc. (A)
11
products. Everything you do in Hoshin is based on data collection and PDCA [Plan-Do-Check-
Act cycle]."
Hoshin also leveraged many of the techniques that ADI learned from its QIP program.
Nevertheless, the company found that implementing Hoshin required more effort than expected.
Hoshin was believed to provide a mechanism for growth, so it assumed a prominent place on the
ADI scorecard. The computer products division placed its Hoshin measures at the top of its 1996
division scorecard; the measures took a key position on scorecards throughout the firm. Hoshin
goals for ADI in 1996 were (1) 98 percent on-time delivery to platinum accounts and (2) 25
percent sales from new products (products introduced within the last six quarters).
KEY SUCCESS FACTORS
Complementing ADI's scorecard were measurements called key success factors that measured
milestones related to the firm’s business plan. These factors were monitor points for the tactical
plans of ADI's strategy. They derived from the company’s five-year plan, which was updated
yearly or, in some cases, quarterly because of the speed at which ADI's market was changing.
The company believed its key success factors were more closely related to wealth creation than
was the scorecard.
One difference between the two measures was that the key success factors were discrete events.
As Suttler stated, "The key success factors do not lend themselves to quarterly monitoring. On
the other hand, the measures on the scorecard are intrinsically limited, failing to capture key
milestones in each business strategy that also need periodic review." The key success factors
either were met or they were not. They did not continue from quarter to quarter. The scorecard
measures, such as on-time delivery, continued to be tracked every quarter.
In 1994 Stata charged Suttler with integrating TQM and planning. Suttler introduced several new
tools, such as the 10-step planning methodology adapted from Hewlett-Packard, to try to
understand wealth creation in a way TQM had not addressed. These techniques were used during
the five-year planning exercise and in developing tactical plans for ADI. The key success factors
were developed with these tools.
Analog Devices, Inc. (A)
12
One aspect of the new planning system was that teams, rather than a centralized planning group,
did the planning. These teams included individual contributors, line managers, and mid-level
managers—the very people responsible for implementing the plans they developed.
Sixty teams were set up in 1997 to work on topics such as business plans and competitor
analysis. ADI management believed the teams would be highly committed to the business plans
because they had developed them.
COMPENSATION SYSTEM
Analog Devices did not link incentive compensation to performance on the scorecard measures.
Compensation for corporate officers (senior management) was based on appreciation in stock
prices. Compensation for all other employees was based on an equally weighted combination of
two factors: growth in company revenues and operating profits for the company before taxes.
Noted Suttler, "We do not tightly link managers' compensation to scorecard performance. We are
in an industry that is moving like the wind. We've got to change our scorecard every year to
respond to that environment. Our compensation philosophy is based on cross-functional
coordination and the highest degree of teamwork."
VISION 2000
Now is the time to set a new vision for the future that builds on the
accomplishments of the past; fully exploits our leadership in signal processing;
captures new opportunities in rapidly emerging markets; and catapults ADI to a
multi-billion dollar enterprise.10
As part of its 1995 planning process, ADI developed what it called Vision 2000 which set forth
three major objectives:
• Build leadership positions in seven critical areas of signal processing.
• Increase the growth rate for sales and profits to greater than 20 percent.
10 Vision 2000: Leadership for the 21st Century, Analog Devices, Inc., 1996.
Analog Devices, Inc. (A)
13
• Grow the organization and develop the skills and competencies of all employees.11
Through the Hoshin process, these goals had been translated into specific and measurable
objectives for every function of ADI. Vision 2000 also included a plan for 1996 that divided the
critical measures into four business drivers (see Exhibit 8).
Each business driver was supported by underlying objectives. For example, the objective of 95
percent on-time delivery by year-end supported the customer satisfaction business driver. The
objectives for the four business drivers for 1996 were clear and measurable. Some of these
objectives, such as on-time delivery, were also part of the scorecard. Others objectives, such as
achieving $175 million in net research and development spending, appeared as key success
factors.
ADI was using Hoshin, QIP, the corporate scorecard, and critical success factors to create,
deploy, and implement strategy. While the systems were in place, the questions still remained:
How should Stata and ADI implement needed change? How important was ADI's corporate
scorecard in creating change? How must these systems evolve for ADI to achieve its objectives
for Vision 2000?
11 Ibid.
Analog Devices, Inc. (A)
14
Exhibit 1: Three Examples of Half-Life12
12 Ray Stata, "Organizational Learning—The Key to Management Innovation," Sloan Management Review, Spring 1989, p. 69.
Analog Devices, Inc. (A)
15
Exhibit 2: Analog Devices Quality Improvement Goals13
Measurement 1987 Half-Life
(in months)
1992
External
On-Time Delivery 85% 9 >99.8%
Outgoing Defect Level 500 PPM 9 <10 PPM
Lead Time 10 weeks 9 < 3 weeks
Internal
Manufacturing Cycle Time 15 weeks 9 4-5 weeks
Process Defect Level 5000 PPM 6 <10 PPM
Yield 20% 9 >50%
Time to Market 36 months 24 6 months
13 Ibid., p. 70.
Analog Devices, Inc. (A)
16
Exhibit 3: Example Corporate Scorecard for FY 1988