2014_01_기술혁신세미나
Session 11
New Product Development 2
OM 박사과정 김수연
OM 석사과정 김인희
OM 석사과정 김경순
Introduction
기업 내부 역량

Manufacturing capability

Integrated Problem Solving

Leadership(PM) & Organization

Flexible Development Process

Development productivity
: overlapping,

Lead time

Product Quality

Cost

Etc
: reward, CAD, …
NPD 성과
기업 외부와의 NW

Supplier Involvement

Market Feedback
2
Product development performance
: strategy, organization, and management in the
world auto industry
Clark, K. B. and T. Fujimoto (1991).
1. Manufacturing Capability
• Elements of successful NPD
: Focus on Design process, Solution of technical problems
: Manufacturing capability(ability to make things rapidly & efficiently) is a critical
sources of advantage in product development
• Manufacturing Productivity & Development Productivity
: positive correlation
4
1. Manufacturing Capability
• Hidden manufacturing activities in the development process
•
1. Prototype Fabrication
Short
(6 mo)
Test engineers can
begin prototype
evaluation earlier
Reasonably
high
More engineering
changes resulting
from tests can be
made in time
Mostly
production
suppliers
Prototype
as problem
detector
Reduce the
negative impact of
late engineering
change
5
1. Manufacturing Capability
• Hidden manufacturing activities in the development process
•
2. Die development
1) Lead time(Performance of die development)
JIT
Philosophy
: Japanese firms have
less WIP dies than US
Supplier
Network
-US
: arm’s-length contract with
separate companies
2) Cost(Performance of die development)
- Japan : at most 20 % of the cost of a die
- US: 30~50 % of the cost of a die
-Japan
: close & long term relationship
with major suppliers
6
1. Manufacturing Capability
• Hidden manufacturing activities in the development process
•
3. Pilot Run & Production Start-up
3.1. Pilot run
- A full-scale rehearsal of the commercial production system
- Objective: to find and solve problems undetected during the prototype production
& testing
Japan
U.S. & Europe
Problem solving cycle
Fast
Slow
Location of pilot
production
Unseparated
(empty hanger approach)
Separated
 Reducing the opportunity cost for the pilot run by minimizing lost production.
 Future workers can get early training in a very realistic setting.
 Workers tend to get excited about a new model they see on the line.
 Workers become motivated to learn more about a new model.
7
1. Manufacturing Capability
• Hidden manufacturing activities in the development process
•
3. Pilot Run & Production Start-up
3.2. Production Start-up(Ramp-up in assembly)
- The start-up of commercial production system
- Objective: same as pilot run
- Effective production ramp-up depends on manufacturing capability and its fit with
choice of ramp-up curve
 Complexity ↓
 Task Continuity ↓
 Task Continuity ↓
 Stable operating condition ↓
 Stable operating
∵ manufacturing capability
condition ↓
U.S. & Europe
(process control)
Japan
Japan
8
2. Integrating Problem Solving Cycle
• Successful product development can be achieved by
: Effective integration of different functional group and engineering discipline
- cross functional team, simultaneous engineering
• 5 Dimensions of Integrated Problem Solving
Limited
communication
Reduce lead time
Rich, dense, early
communication
9
2. Integrating Problem Solving Cycle
• Successful integration problem solving can be achieved by
: Stage overlapping (high degree of simultaneous activity)
: Intensive Communication (rich, frequent, bidirectional information flow)
10
2. Integrating Problem Solving Cycle
• Integrated Problem Solving & Development Performance
1. Integrated Problem Solving
- “ideal profile index” composed by 9 characteristics
(ex: high simultaneity ratio, early feedback from manufacturing, etc)
2. Development Performance
- lead time, development productivity, total product quality
Japan Effect
Positive correlation
11
3. Leadership and Organization
• 조직의 형태와 리더십은 제품개발성과에 영향 미치는 중요한 요소임
• 3 Dimensions of Product Development Organization
Organizational
dimension
Specialized
Expected function
• Accumulate and preserve technological expertise at the level of
individual components and activities
• execute individual tasks with speed and efficiency
Internal Integration
External Integration
• Achieve high internal integrity of total product
• Achieve fast product development through better task coordination
• Achieve high external integrity of total product
• Match product concept, product design, and customer expectation
12
3. Leadership and Organization
• 4 Modes of Development Organization
• Broader responsibility and clout.
: internal/external coordination
: product planning & concept development
13
3. Leadership and Organization
• Patterns of organization and leadership
- Measure
•
Specialization: Number of people who participate in a project on a long-term basis
•
Internal Integration: Engineering coordination indicator (based on PM role)
•
External integration: Concept development & Market link indicator (Based on PM role)
•
Performance: Development Productivity, Adjusted lead time, TPQ Index
Specialization
U.S
Internal
Integration
Japan
External
Integration
Japan
14
3. Leadership and Organization
• Organization & Performance
 Shaded diagram refer
to high significant of
relationships(.03)
 External integration, in
contrast, is closely
associated with total
product quality but not
with lead time or
productivity
15
3. Leadership and Organization
• Organization & Performance
 Lower performers generally lack integration.
 Some Japanese firms employing the middle
weight system are both fast and efficient.
 The very different circumstances of high-end
specialists were acknowledged earlier, but even
volume producers occasionally spawn light
weight pm achieving solid results in a single
dimension of performance(TQM).
16
Discussion Questions
1. Dimension of integrated problem solving에서 우측으로 갈수록 lead time을 줄일 수
있는 좋은 problem solving form이라고 하였다. 이에 대한 반대 의견은?
2. 본 연구에서 제시한 5가지 dimension 외에 추가할 수 있는 dimension에는 무엇이 있을까?
Limited
communication
Reduce lead time
Rich, dense, early
communication
17
Discussion Questions
3. 본 연구에서는 신제품 개발 조직 구조가 functional structure일 때보다 heavyweight
product manager일 때, 더 좋은 성과를 가져온다고 하였다. Heavyweight product
manager를 가져가는 조직 구조의 단점은 무엇이 있을까?
18
Developing product on “Internet Time”
: The Anatomy of a Flexible Development Process
Alan MacCormack
Roberto Verganti
Marco Iansiti
(2001)
1. Research Objective
1. Many literatures on the Effective Development process before this article
2. Previous authors have developed theoretical models which demonstrate
the value of greater flexibility in a design process faced with uncertainty.
3. This study attempts to examine the underlying mechanisms through
which firms directly influence the flexibility of their development
processes
20
2. Models of the Product Development Process
Figure 1 - A stage-Gate Model of Product Development (Traditional Water fall model)
 The product concept is defined and frozen prior to the start of
detailed design, and the functionality this specifies in each module is
completed prior to the start of system level test.
 The First challenge this model faces is that it assumes all information
about potential design choices is known or can be discovered during
concept development.
 Second challenge is that feedback on how the product performs as a
system is not obtained until late in a project, when the functionality
in each module has been fully developed
21
2. Models of the Product Development Process
Figure 2 - A More Flexible Model of Product Development
 Recent studies have began to investigate more flexible models of
development characterized by the overlapping of development stages
(e.g. Krishnan et al. 1997).
 In this model, development becomes an “evolutionary” process of
learning and adaptation (Tushman and O’Reilly 1997).
 This model commonly used for software development.
 These models have been proposed to address flaws in the traditional
waterfall model of development.
22
3. Hypotheses
 These two objectives (i.e. maximizing product performance and facilitating process flexibility) are
often incompatible from a design standpoint; hence, the selection of the “optimal” product
architecture becomes a more complex problem (Ulrich 1995).
Hypothesis 1. In uncertain and dynamic environments, greater investments
in architectural design will be associated with better performing projects.
 Where information about the user environment is tacit or “sticky” (i.e., not easily captured by
traditional market research techniques) there is value in mechanisms which facilitate the release of
early product versions to users when a project retains the ability to change the design.
Hypothesis 2. In uncertain and dynamic environments, earlier feedback on a
product’s system-level performance will be associated with better
performing projects.
 Lower-level learning results in knowledge that can be directly applied to a specific context,
whereas higher-level learning results in a deeper knowledge of the process of problem solving,
especially with respect to analyzing new frames of reference.
Hypothesis 3. In uncertain and dynamic environments, development teams
with greater amounts of generational experience will be associated with
better-performing projects.
23
4. Measures
• Sample: 29 completed projects(unit of analysis) from 17 firms in internet
software industry.
• Performance(Dependent variable): Product Quality. Overall quality was
defined as a combination of product features. Use the assessments of a
panel of experts gathered using a two-round Delphi process.
• Control variable: Resources. Resources allocated to each project to control
for the potential impact of resources on product quality.
• Development Process(Independent variables)
• Investments in Architectural Design: The ratio of architecture design
resources to development and test resources.
• Early Market and Technical Feedback: Percentage of the product’s
functionality that has been developed when each milestone(prototype
during the concept development stage, system integration, the first
beta release point) is reached.
• Generational Experience: the proportion of team members with
greater than two generations of experience.
24
5. Results
25
5. Results
Feedback 중 Tech Feed back의 경우 프로토타입과 베타테스트와 상관
관계가 높게 나와 제외하고 분석을 시행함.
26
5. Results
상관분석 결과 Control 변수로 사용한 Resource와 경험 변수가 상관관계
가 높은 것으로 나왔음.  구성원의 경험이 Performance에 영향이 없는
것이 아님. 자원의 효율적 배분에 크게 영향을 끼치는 것으로 해석할 수
있음
27
Discussion Questions
•
NPD 과정에 Consumer를 참여시키는 것이 오늘날 trend이다.
Supplier 및 Consumer와의 전략적 관계를 통한 NPD가 가져올 수 있
는 이득은 무엇이 있을까?
Source: M. A. Kaulio(1998), TQM
28
Discussion Questions
Successful Customer involvement Case
LDD = Lego Digital Designer  Design by customer.
신제품 개발의 혁신 동력을 다양한 User로 부터 효율적으로 확보
29
Project Scope and Project Performance
: The effect of parts strategy and supplier involvement
on product development
Kim B. Clark(1989)
1. Introduction
– The importance of product development has motivated significant
attention to the determinants of performance by both practitioners
and academics.
– This study examines one aspect of project strategy, project scope and
its effect on project performance(Lead time and project manhour).
31
1. Introduction
•
Project Scope
– The extent to which a new product is based on unique parts
developed in-house.
– Unique parts have been selected, the firm may rely on a supplier for
engineering work.
•
Two elements of scope
– The choice of unique versus off-the-shelf parts
– Choice of supplier involvement
– Once unique parts have been selected the firm may rely on a
supplier for engineering work, reducing internal engineering efforts
in the project.
32
2. Measure
•
Scope and Engineering Manhours
– Decisions about scope have a direct impact on observed manhours
in the project
NH = 1- b[C+S(1-C)]
 C: Fraction of parts that are off-the-shelf
 S: Fraction of engineering effort for unique parts done by suppliers
 b: Fraction of total engineering effort that is part specific
NH = 1 – (b*C) – [b*s(1-C)]
 b*c: Fraction of off-the-shelf parts
 b*s(1-C): Fraction of unique parts developed by supplier
33
3. Results
Summary data on performance and content by region
34
3. Results
• Scope and Engineering
Manhours (regression analysis)
• Adding measures of
project scope has a
dramatic effect on the
regression.
• Interaction between price
and NH  imply that the
impact of scope on
manhours depends on the
complexity of the product.
• Japanese dummy excluded
from the regression: the
coefficient on the supplier
variable increases by a
factor of 2.7  quality of
the relationship and the
way that it is managed is
important.
35
3. Measure
•
Scope and Lead time
– Impact of scope on lead time is not straight forward.
– Lead time: Time elapsed between start of the development project
and market introduction.
– Lead time is determined by the critical path in the network
– Supplier involvement may reduce lead time if a supplier were more
capable in executing parts engineering
36
3. Results
Scope and Lead Time (regression
analysis)
• Strong positive impact of scope on
lead time
 increase in scope from the Japanese
level of 0.57 to the U.S. level of 0.66
would increase lead time by 3.9
months
• The effect of Interaction between
price and NH is insignificant
 scope on lead time does not
depend on the complexity of the
product.
• Japanese firms derive real advantages
from their supply base.
37
3. Results
• Who does the work, and how the content gets implemented (offthe-shelf vs unique parts), makes a difference in the length of
the planning process.
38
Discussion Questions
• 기업이 NPD에 Supplier를 참여시키고, 이들의 역량을 적극적으로
활용하는 것이 항상 긍정적인 영향만을 가져다 줄까? Supplier의
참여로 발생하는 문제는 어떤 것이 있을까?
39
Discussion Questions
 Executives, viewing manufacturing mai
nly as a cost center, give shot shift to
the impact that outsourcing or offshor
ing it may have on a company’s capac
ity to innovate.
 Indeed, most don’t consider manufact
uring to be part of a company’s innov
ation system at all.
 Massive migration of manufacturing fr
om United States has seriously eroded
the domestic capabilities needed to tu
rn inventions into high-quality, cost-co
mpetitive products, damaging Americ
a’s ability to retain a lead in many sect
ors.
40
41
Discussion Questions
• The problems In managing supplier Involvement
(FINN WYNSTRA et al. 2001)
• The relation as a source of problems
• The supplier as a source of problems
• The manufacturer as a source of problems
• Solution
• Identifying specific processes and tasks that need to
be carried out, aimed at the integration of product
development and sourcing processes;
• Forming an organization that supports the execution
of such tasks; and, finally,
• Staffing the organization with people that have
42
Accelerating Adaptive Processes
: Product Innovation in the Global Computer
Industry
Kathleen M. Eisenhardt
Behnam N. Tabrizi(1995)
1. Comparison of compression and experiential model
Compression Model
Experiential Model
Key Assumption
Predictable or Certainty
Unpredictable or Uncertainty
Image of product
innovation
Predictable series of well-defined
steps
Uncertain path through foggy &
shifting markets & technologies
Strategy for speed
Rationalize & then squeeze the
process
Quickly build understanding &
options while maintaining focus &
motivation
Bureaucratic
Flexible
Organization features
Hypothesis
(Independent Variables)
Dependent Variables
Data and Method
Routine
(H1)
(H2)
(H3)
(H4)
(H5)
(H6)
Planning
Supplier Involvement
Using the CAD
Overlap
Multifunctional team
Greater Award
Improvisational
Experimental
(H7) Multiple Iteration
(H8) Extensive Testing
(H9) Frequent Milestones
(H10) Power Leader
Development Time
Survey and deep interview results from 72 projects(36 Computer
Companies)
44
2. Compression model
H1: More time spent in planning is associated with shorter development
time.
H2: More supplier involvement is associated with shorter development
time.
H3: More designers using computer-aided design(CAD) is associated
with shorter development time.
H4: A higher degree of project overlap is associated with shorter
development time.
H5: Multifunctional teams are associated with shorter development
time.
H6: Greater reward for schedule attainment is associated with shorter
development time.
45
3. Experiential model
H7: More design iterations are associated with shorter development
time.
H8: More time spent in testing throughout the development process is
associated with shorter development time.
H9: Less time between milestones is associated with shorter
development time.
H10: Greater power of the project leader is associated with shorter
development time.
46
4. Methods
-
Data from 72 product development projects drawn from European, Asian, and
U.S. computer firms.
-
Computer companies that compete in the personal computer(PC), minicomputer,
mainframe, and peripherals segments of the industry.
-
Unit of analysis: product development project
-
Within each company, two product development projects within a single product
group were used.
47
5. Result
Opposite
Compression
Model
Opposite
Opposite
Experiential
Model
Compression
Model
Experiential
Model
Compression
+ Experiential
Parsimonious
model
48
5. Result
•
Split-sample Analyses
Certain products
49
5. Result
•
Split-sample Analyses
Less Certain products
50
6. Discussion Question
• 본 연구에서 제시되는 Independent variables들의
concept이 올바르게 측정되었다고 생각하는가?
construct
Planning
Reward for
schedule
attainment
total reported time spent/total
elapsed time
Planning:
 Extensive planning
simply waste time,
especially high-velocity
industry.
Asking whether the product
development personnel were
rewarded for meeting the
schedule deadlines during the
course of the project
Reward:
 reverse causality: slow
firms may use reward
for schedule as a
solution to their pace
problem.
measure
 Neglect quality
problem.
51
Compression Model-IV
Construct
Measure
Supplier
Involvement
Numbers of stages that suppliers participated
CAD(Computeraided design
system)
1) Percentage of design engineers on the product
development team who used CAD
2) Rating by respondents of the importance of
increasing CAD usage to decreasing product
development time
Project overlap
Sum or overlaps across the six processes of the product
development/total development time
Multifunctional
team
Adding the total number of functions in every phases
(Whether personnel from specific functions were
involved on the product development team during the
various stages of product development)
52
Experiential Model-IV
Construct
Measure
Number of
design
iterations
- Asking the number of design iterations for a project.
- Adjusted by the average number of iterations for each
industry segmentation.
Test
Reported time spent on testing/total elapsed time
Time
between
milestones
- Average time between milestones during the project
- Adjusted by the average time between milestones for
each industry segmentation
Power of the 1)Asking respondent group where in the SBU hierarch the
project leader project manager reported(1,0)
2)Assessed power by whether or not the project manager
was the final decision maker on the key issues of budget,
team composition, and project timetable.
53
Dependent and Control Variables
Construct
Measure
Development
time
-Duration: start and end da
tes of the project
-Development time: adjuste
d by average project durati
on of its corresponding ind
ustry segment
Project size
Dummy variable
0: medium-scale project
1: major project
Development time:
-start of a project:
1)the time at which the first meeting
was held to consider the development
of this specific product.
2)First assignment of firm personnel to
assess the feasibility of the product.
-ending date: the date at which
product stabilization was reached.
*stabilization: the time at which no
more changes are made to the
product
Project size:
Major projects would take longer to
complete than medium-scale ones.
54
6. Discussion Question
• 본 연구의 실증분석 방법에 대해 어떻게
생각하는가?
 Compression/Experiential strategy’s assumption
: industry, product characteristics, degree of maturity
55
6. Discussion Question
• 본 연구에서 밝혀낸 IV(multifunctional team,
multiple iterations, extensive testing, frequent
milestones, powerful leader)들이
제품개발시간을 줄이는 것을 방해하는 요소에는
어떤 것들이 있을까?
56
Woodman, Sawyer, and Griffin(1993)
57