Summary of methods and results for reducing cost, driving quality upstream, optimizing systems, managing suppliers, accelerating time to market, and improving performance

1. +
Tim Rodgers
Methods and Results

2. +
Method: Design for Low Cost
The design
is what it is,
we’re stuck
with it
Typical approach, design is the constraint
• Lower cost materials
• Lower cost suppliers
There are other
designs that
• Lower cost factories
meet customer
requirements at
lower cost
Re-design, challenging the assumptions:
• DFX: manufacturability, sourcing, reliability, support
• Integration: one part or subsystem doing multiple jobs
• Simplify, remove low-use / low-value features
• Reduce size and weight for lower shipping cost per unit

3. +
Method: DFM for Printed Circuits
Designer
interviews re:
common
tradeoffs
DFM
Manual
Tools that enable circuit
designers to choose options
based on the relative cost in
materials and processing
Production
cost & yield
data
Example
More circuit layers
(material cost)
Smaller conductors
(yield loss)

4. +
Method: Analyzing Cost of Quality
Field repair, customer
support, and other warranty
costs
Design failure:
Design doesn’t meet
requirements or isn’t
robust under a range of
operating conditions
(often appears as a part
failure)
Tolerance failure:
Design fails to
account for natural
variation in part
characteristics and
assembly processes
Part
design
System
design
Internal yield loss,
scrap, and rework
Production
process
design
Production process
failure:
Improperly assembled
from good parts, or
damaged prior to
shipment
Work
instructions
& training
Part failure:
Part did not meet
the performance
required by the
design
Supplier
performance

5. +
Results: Lower Cost Production
Monthly
Net Yield
Target
Jun-2010
Jul-2010
Aug-2010
Sep-2010 Oct-2010
Nov-2010
Dec-2010
Jan-2011
88.02%
91.80%
92.87%
93.18%
94.81%
95.49%
96.65%
97.01%
96.95%
95.00%
95.00%
95.00%
95.00%
95.00%
95.00%
95.00%
95.00%
95.00%
yield
Net Yi el d
98.00%
94.81%
96.00%
92.87%
94.00%
95.49%
96.65%
97.01%
91.80%
Design change to
provide greater
assembly tolerance
90.00%
88.02%
86.00%
84.00%
Mfg process change to
reduce defects for
critical subassembly
82.00%
Tar get
93.18%
92.00%
88.00%
Feb-2011
Assembly jigs to reduce
variability during part
installation
96.95%
Better ESD protection
for critical
PCAs, eliminating
accidental discharge
during assembly
Increased production capacity by 10%
Reduced operating expenses per line
Estimated savings = US$1.1M per year
80.00%
Jun-2010
Jul-2010
Aug-2010
Sep-2010
Oct-2010
Nov-2010
Dec-2010
Jan-2011
Feb-2011

6. +
Results: Lower Cost Factory Test
Top-Level Assembly
100% end-of-line
testing
Rework
Improved yield
Fewer tests and consumables
Product quality &
reliability audit
testing
Packaging and
outbound audit
Reduced sampling for audits
Eliminated rework stations
Lower operating expenses
Faster throughput

7. +
Results: Lower Cost SW Testing
Before
• Many test
cases, routinely
repeated for each
SW release without
regard to the causes
of call rates and
warranty cost
• No test design
process focused on
high-risk modules
and interfaces
• US-based test
execution, low job
satisfaction, high
turnover
Actions
• US-based team
trained and reassigned to test
case development
and project
management
• Test development
based on
collaboration with
design team and
feedback from
customers
• Test execution
outsourced to
offshore service
providers
Benefits
• 80% lower test cost
per development
project as a result of
offshore testing and
fewer (but focused)
tests
• 30% lower customer
call rates as a result
of test engineering
• Higher job
satisfaction, lower
turnover

8. +
Results: Lower Support Cost
Prevention
Rapid Deployment
Field performance and internal
verification prior to release
Sustaining engineering managed
as separate projects
Warranty cost
reduced by
$300K per year
Prioritization
Design for Support
Focus sustaining engineering on
the leading contributors
Modular design with a minimum
number of unique configurations

9. +
Method: Support for Quality
Assess the
competitive
environment
Determine the
risk to future
business
Measure the
cost of quality
Crossfunctional
support for
quality
initiatives
> Internal =
scrap, rework, hours spent
managing quality issues
> External = warranty
cost, field repair

10. +
Method: Driving Quality Upstream
Upstream in the product development process
Design-in quality and verify before ramp
Design
Factory
Product
Suppliers
Upstream in the value delivery system
Hold suppliers accountable for quality
Cost to
address
quality
issues

11. +
Method: Quality Maturity Model
Prevention based on proactive
analysis of the design
Increasing Cost
Effectiveness
Control parameters that are critical
to product performance, out-of-box
quality, and reliability
Analyze failures to understand causes
Improve design, part quality, and
production processes to make failures
less likely
Internal failures
Improve test & inspection
Corrective action
External failures
Corrective action

12. +
Method: SW Development & Test
Checkpoint
Development Priorities
Role of Testing
Invention & investigation
New features & capabilities
Rapid, focused
evaluation of new
functionality
“Does this work?”
Design improvements
Additional robustness
Broad and deep: many use
cases and configurations
“Find defects so we can fix
them”
Optimize for final release
Emphasis on defect fixes
Rapid, focused regression to
verify defect fixes
“Did we fix it without breaking
something else?”
Functionally
Complete
Code Freeze
Final Release

13. Results: Rework from China Factory
Problem:
Excessive rework on
subassemblies from China
factory
Rework cost
Possible solution:
Outbound inspection
$20,000
$18,000
$16,000
$14,000
$12,000
$10,000
$8,000
$6,000
$4,000
$2,000
$0
W17
W18
W19
W20
W21
W22
W23
W24
W25
W26
W27
W28
W29
W30
W31
W32
W33
W34
W35
W36
W37
W38
W39
W40
+
Better solution:
 Raising awareness of costs
with local managers
 Training
 Team incentives
 Work instructions in local
language instead of English

14. +
Results: Software Install Time
Original SW installation process



Focused design effort to shorten
time and eliminate defects
Up to 30 minutes to install on a
single system
25%
Confusing process including
specific manual steps that must be
executed in sequence
20%
Many opportunities for the user to
interrupt the process, requiring a
lengthy phone call to customer
support
40
35
30
25
15%
20
10%
15
10
5%

Customer less likely to use the SW
after installation, negative impact to
revenue and loyalty
Call rate
5
0%
0
Before
After
Average
length of call
(min)

15. Results: Part Quality Improvement
4000
3500
Leadership of kaizen projects
at this critical supplier reduced
inspection and rework costs
by 45%
3000
Defect PPM
+
2500
2000
1500
1000
500
0
Goal

16. Results: Eliminating Inspection
100
99.8
Net EOL Yield (%)
+
99.6
99.4
99.2
99
98.8
Target = 98.5%
Stable platform consistently
exceeding customer’s quality
goals
98.6
98.4
98.2
98
Excessive inspection reduced
margins
 Eliminated all in-process inspection, saving 5 people per shift
 Reduced incoming part sampling rate for most parts, and reverted cost of
remaining incoming inspection to suppliers, saving 8 people per shift
 Implemented SPC on critical factory processes to provide earlier detection
of quality issues and control

17. +
Method: Process Investigations
Process is not
being used (people
problem)
1. Do they know there is a process?
Inform them
Process is
inefficient or
ineffective (process
problem)
2. Do they know how to use the process?
Train them
Does the process take too long?
3. Do they ignore the process?
Apply Theory of Constraints
Explain to them
(identify and manage the bottleneck)
4. Do they have a better way?
Learn from them
Does the process fail to deliver expected results?
Check output from each step
Does the process deliver inconsistent results?
Determine the causes of variability and eliminate

18. +
Method: Process Convergence
Switching cost
Cost to support multiple
versions of the process
Cost due to incompatibility
Clearly
articulate
expected
benefits
Change
management
Communicate
value of
convergence
Determine
value of local
configurations

19. +
Results: Factory Line-Down
Process
Describe
Current
Situation
(Is / Is
Not)
Propose
Solutions
4
Propose
Possible
Root
Causes
Test
Solutions
Define
Problem
Integrate
and
Monitor
Change
Test Root
Causes
100%
80%
3
60%
2
40%
1
20%
0
0%
Before
After
Cycle Time
(hr)
Recurrence
(%)
After:
Systematic approach
Longer process, but highly effective
Lower overall cost

20. +
Results: Design Change Process
Situation: Part design changes to
reduce material cost or address field
quality issues hampered by a slow
and ineffective process
Slow approval &
implementation
through the supply
chain
Lack of coordination
Rapid deployment
to realize cost and
quality benefits
Removed unnecessary steps
Blocked “jumping” over required
steps
Added auto notification for steps
taking longer than standard time
Improved software tracking tool

21. +
Method: WW Supplier Experience
Germany
France
UK
Ireland
Spain
Hungary
Czech
Rep.
US
Canada
Mexico
Brazil
China
Japan
S. Korea
Taiwan
India
Singapore
Malaysia
Thailand
Indonesia

22. +
Method: Supplier Audit Program
Competitive
quote from
RFQ
First articles
pass
inspection
Is the supplier capable of
sustaining performance?
Routine Audit:
• Management commitment
• Statistical process control
• Problem solving
• Incoming inspection
• Training, work instructions
• Preventive maintenance, calibration
• Specifications and document control
• Internal audits
• Record keeping
• Shop floor control & 5S

23. +
Method: Supplier Quality Maturity
DFM feedback from
understanding of
design requirements
Proactive warning of
supply issues
Preventive action to eliminate
root cause
Process management and
control
Less testing & inspection
Basic performance
Meets requirements
Rapid corrective action as issues are reported

24. +
Method: Driving Quality Upstream
Product characteristic
(e.g., functionally critical
dimension)
Production process
parameter (leading
indicator)
Process capability to
consistently meet
specification
Process control to
reduce variability
It’s not about
preventing bad parts
from being shipped …
It’s about preventing
bad parts from being
built in the first place

25. +
Method: Benefits of Partnership
Advertised relationship with well-known customer
(especially valuable for small suppliers)
Predictable demand for better asset utilization
Contractual commitment to fixed capacity
Technical capabilities that can be leveraged to
other customers
Unique market that provides balanced portfolio
Loss of business (balanced by
customer’s switching costs)

26. +
Results: Favored Supplier Program
Favored suppliers (based on quality performance)
Favorable pricing and payment terms
Low inventory, ship-to-stock
Accelerated qualification of new part numbers
Audit inspection of incoming lots
Other suppliers
Incoming inspection charges reverted
(First article failures, defective parts found on the line)

27. Results: Molded Plastic Supplier
Supplier provided large quantities of
injection molded plastic parts
Problem: inability to consistently meet
critical dimensions and cosmetic
requirements
Flash
Pressure
+
Shorts
Melt temperature
Optimum temperature and pressure had
been defined for the part, but the supplier
had failed to conduct a window study to
determine the allowed ranges, or account for
mold wear and cavity variations
Performance improved after the supplier
established process control limits and regular
sampling from cavities with support from the
customer.

28. +
Results: Gold Plating Thickness
Supplier provided gold plated contacts for
printed circuits.
Problem: gold thickness varied outside the
spec limits
Control chart for gold thickness indicated a
process that varied outside 3 sigma limits.
Concentration of gold in the plating tank was
not monitored regularly and chemical
additions were made based on rough
estimates.
With support from the customer, the supplier
implemented a regular laboratory analysis
and strict controls on chemical additions.

29. +
Method: Rapid Time-To-Market
Invention & investigation
Experiments, feasibility studies
Evaluate concepts vs. requirements
Converge on a product design
Execution
Design verification:
Production system verification:
Can you build one unit that meets
requirements?
(rapid prototyping)
Can you build many units that meet
requirements?
(early engagement with supply chain)

30. +
Method: Re-Use and Leverage
Product Development
Model
Advantages
Limitations
Single product: Development
of a single product, feature
set and price point: one-at-atime to respond to the market
Laser-focus, enabling design
to achieve minimum cost for
that feature set without being
constrained by earlier
choices
Limited design re-use,
possibly lengthening the
development time; design
costs, tooling and parts not
amortized across a larger
number of units
Platform: Development of a
fixed core or foundation that
becomes the leveraged,
common basis for follow-on
derivative products
Initial design and tooling
investment is leveraged to
subsequent derivatives or
extensions with rapid time-tomarket
The initial platform design
limits derivative product
design options, reducing
flexibility and market
responsiveness
Architecture: Development of
a set of interchangeable
modules or assets with welldefined interfaces that are
designed to enable
forecasted product options to
meet anticipated needs
High level of design flexibility;
initial investment enables
lower product development
cost for later products;
leveraged tooling and faster
time-to-market; can easily
shift to other development
models
Modules are optimized for
leverage and re-use, not
necessarily for absolute
minimum cost

31. +
Method: Development Phase Gates
• Is the design capable and stable?
• Indicators: critical performance requirements met; design turmoil;
open issues from FMEA and other risk analysis; prototype defects
Design Gate attributed to design; design margin study complete
Process
Gate
Production
Gate
• Are the production processes (including the supply chain)
capable and stable?
• Indicators: production and test yields; prototype defects attributed to
WMS and material; changes to tools, fixtures & work instructions
• Is the factory ready to build at full production volume?
• Indicators: no open waivers (all resolved or closed); all tools & fixtures
meeting GRR requirements; functionally critical dimensions and
parameters meeting process capability requirements

32. +
Method: SW Development Checkpoints
Traditional “waterfall” software delivery model
Test &
Defect
Fixing
Development
Fixed
release
date
o Testing begins after development is “complete”
o Scramble to find and fix high priority defects
o Quality usually sacrificed to meet schedule
Phased “prototype” model based on hardware design
Development
Modular
architecture, core
functionality verified by
design team before
check-in
Functionally
complete
checkpoint
System
Testing
Final
Test
Code freeze
checkpoint
Defect fixing, additional noncritical enhancements
added and verified
Lower cost development, better quality
Later replaced by
agile/scrum with
staged development
& testing

33. Method: Time to Achieve Quality
Target
Quality
+
Failure to meet
quality target at
start of
production
#1
#2
#3
Prototype builds
Start of
production
Additional cost
and loss of
production
capacity

34. Results: Stable Design at Ramp
100
100
99.5
98
99
96
94
92
Target = 95%
90
88
Net EOL Yield (%)
Net EOL Yield (%)
+
98.5
98
97.5
Target = 98.5%
97
96.5
96
95.5
95
Insufficient attention to DFM and
quality during development
 Improvement after ramp required
repeated problem solving to determine
root causes and successfully eliminate
them
 Higher cost and delayed product
introduction until issues resolved
 Emphasis on design stability
 No design changes permitted after last
manufacturing readiness build
 Steady reduction in design-related defects
throughout the development phase
 Zero open waivers at ramp
 Daily tracking of yield and defects during
prototype builds

35. +
Results: Software Program Tracking
Metrics for tracking status:
 Requirements verified by testing
 Code turmoil (new or changed LOC)
 Remaining open defects (weighted)
 Defect find/fix rate
100%
80%
60%
40%
20%
0%
Tests passed
Tests failed
Tests not run
Tests blocked
Clear understanding of work
remaining to trade schedule
vs. scope vs. quality
Predictable outcomes that
meet business objectives

36. +
Method: Metrics Alignment
Performance
measures
Link business objectives to individual/team
objectives and performance measures
Considerations:
Control over outcomes and improvement
Behaviors that are encouraged
Department
& Team
Objectives
Business
Objectives

37. +
Method: Improvement Cycle
10
8
6
4
2
Improvement Plan
Owner
Date
0
Pareto of Root Causes
1.
2.
3.
4.
12
10
8
6
4
2
0
A
B
C
D
E
F
Measure performance
Identify negative trends
Determine root causes
Develop improvement
plans to address leading
causes
5. Hold owners accountable
for improvement
6. Measure, verify, repeat

38. +
Method: SW Product KPIs
Product
Development
Objectives
Key Performance Indicator
(KPI)
Key processes and behaviors
Value-added: new
features that meet
customer needs
Incremental revenue dollar
contribution as a result of the
new features
Better understanding of customer
needs to target high-value features
On-time delivery
Actual checkpoints vs. plan
Disciplined development
processes, an architecture that
supports design of incremental
features, and high quality to
ensure customer rapid qualification
Deployment
Penetration, percentage of
the installed base using the
new features
Tight collaboration with sales and
field support to drive new orders
and installations
Quality
Escapes, defects reported by
customers, especially those
that prevent the customer
from qualifying and deploying
value-added features
Dedication to zero defects and
reduction of variability

39. +
Results: Cost Based Metrics
Shifting to a cost measure
focuses attention on the
opportunity for savings
The data is accurate, but
doesn’t inspire action
Production Yield
85%
80%
75%
70%
65%
60%
55%
50%
Cost of Quality
$6,000
$5,000
Jan
Feb
Mar
Apr
May
Jun
$4,000
$3,000
Defects per Unit
Scrap cost
Rework cost
$2,000
1.00
0.80
$1,000
0.60
$0
Jan Feb Mar Apr May Jun
0.40
0.20
0.00
Jan
Feb
Mar
Apr
May
Jun

40. +
Results: Factory Quality Team
Function
Primary Responsibilities
Key Performance Indicators (KPIs)
Quality program
management
Quality issue
management, primary
customer contact window, data
reporting
1.
2.
3.
On-time and accurate data reporting
Rapid response and closure of quality issues
Effectiveness of issue management
New product
quality
Quality plan development and
implementation for new
products
1.
2.
Time-to-quality after production ramp
NPI Product Quality Scorecard
Manufacturing
test
Software support, script
development, implementation
and maintenance
1.
2.
False positive or negatives due to test script
Rapid ramp to required level of engineering
capability and independence
Contribution to productivity and COQ
3.
Document
control
BOM and change request
management, other document
control processes
1.
Quality
engineering
Proactive opportunities for
yield, quality and COQ
improvement
1.
Measurable yield, outbound quality, or COQ
improvement
Quality
assurance
Inspection, testing and data
reporting for production lines
1.
2.
Escapes found by subsequent audits
Contribution to productivity
Supplier quality
management
IQC inspection, management
of corrective action with
suppliers, proactive monitoring
of part quality
1.
2.
3.
COQ for IQC
Percent of suppliers achieving favored status
Escapes and line-downs due to part quality
2.
3.
BOM accurate and up-to-date, all partners
informed about changes and revisions
No mistakes due to incorrect PN revisions
Cycle time for change process, from initial
request to approval and implementation
What’s
important is
that each
function has
independent
control over
their KPIs

41. +
Results: R&D Department Aligned
Every activity should
contribute, or why do
it?
New feature
Sales, revenue
Improved performance
Operating expense
Improved cost or quality
Productivity, throughput
Improved infrastructure
Gross margin /
R&D expense

42. +
Results: SW Testing Effectiveness
Offshore software testing service engaged as a partner
With encouragement from the customer, the supplier invested in training to
improve testing and increase the percentage of defect reports that
resulted to a code change (defect fix)
The improved quality of service contributed to the steady reduction in
support call rates for each software release
80%
60%
40%
Reported Defects
Fixed
20%
Call Rate
0%
A B C D E F G H J
SW release

43. +
Method: Quality Culture Transformation
FROM
 Passive reporting of quality
issues
 Waiting to react to customer
escalations
 Corrective action to fix the
problem
 Issue closed when plan is
implemented
 End-of-line quality measures
based on testing and inspection
 IQC, sorting, testing, audits,
inspection
 Quality metrics required by the
customer
 Test plans developed and
provided by the customer
 Quality is the responsibility of the
Quality department
TO
 Leadership to closequality issues
 Proactivequality improvements
based on understanding
 Understand and eliminateroot
cause
 Issue closed when improvements
are measured
 In-process measures as early
indicators(SPC)
 Drive quality upstream(design
and parts)
 Cost of quality (COQ)and other
internal metrics
 Quality plans developed with the
customer in mind
 Quality culture in the entire
organization