CMMI and Six Sigma

1
CMMI and Six Sigma

• SC-SPIN
• February 28,2008

2
Presentation Outline

• Introductory Comments
• Characteristics of High Maturity Organizations
• CMMI and Value Streams
• Measuring
• Six Sigma

3
Capability Maturity Modeling Integration (CMMI)

• Develop/Document Process
• Refine Process
• LSS is a tool to assist w/ process refinement
• Standardize process across multiple areas
• Measure the process
• Modify based on measurement analysis

4
Characteristics of High Maturity Organizations

• A mature ML 4-5 organization can accurately,
  quantitatively estimate quality and process
  performance, make confident commitments to those
  predictions, and execute those commitments as
  planned.
• Statistical and other quantitative methods are
  used, at both the organizational and project
  levels, to understand the past, control the
  present, and predict future quality and process
  performance.
• Organizations establish quantitative objectives
  for quality and process performance based on
  their business objectives.
• Individual projects establish their objectives
  based on those of the organization and the needs
  of their customers and end users.

5
Characteristics of High Maturity Organizations

• Projects and individuals use statistical and
  other quantitative methods to plan, monitor, take
  corrective actions, and predict progress against
  their objectives.
• Key subprocesses are identified and statistically
  managed.
• The performance of key subprocesses is known and
  used to compose a overall processes to meet
  organizational and project objectives.
• High Maturity capabilities allow organizations
  to
• Use information from individual projects to
  understand and refine project performance and
  variation for the organizations standard
  processes
• Know what their process limitations are so they
  do not over commit
• Target areas for continuing improvement
• Evaluate quantitatively the impact of proposed
  improvements.

6
Characteristics of High Maturity Organizations

• Prerequisites for a High Maturity organization
• The ability to gather and use meaningful data at
  all levels from individual practitioners to
  projects to the organization itself
• Defined processes for projects that specify how
  and when data are collected for use in
  quantitatively managing the projects
• Consistent tailoring of the projects defined
  processes from the organizations standard
  processes
• A functioning Measurement and Analysis (MA)
  process area that collects data at the
  practitioner and project levels and elevates that
  data to the Organizational Measurement Repository
  for use by other projects
• An understanding of what constitutes meaningful
  and useful measurement data
• An understanding of the types of variation that
  are present in all processes

7
CMMI Maturity Level 4 Process Areas

• Organizational Process Performance (OPP)
• The purpose of Organizational Process Performance
  (OPP) is to establish and maintain a quantitative
  understanding of the performance of the
  organizations set of standard processes in
  support of quality and process-performance
  objectives, and to provide the process-performance
  data, baselines, and models to quantitatively
  manage the organizations projects.
• Quantitative Project Management (QPM)
• The purpose of Quantitative Project Management
  (QPM) is to quantitatively manage the projects
  defined process to achieve the projects
  established quality and process-performance
  objectives.
• OPP and QPM are very tightly coupled and require
  very close cooperation between the organization
  and the projects to realize their benefits.

8
CMMI Maturity Level 4 Process Areas

• Comments on Organizational Process Performance
  (OPP)
• The organization builds a set of Common Measures
  (in the Organizational Measurement Repository)
  that represents actual performance of processes
  for individual projects. These Common Measures
  are statistically analyzed as to distribution and
  range and then applied to any individual project
  in the organization.
• Quantitative objectives for quality and process
  performance are established for the organization
  and are based on the organizations business
  objectives and the actual past performance of
  projects.
• Performance Baselines and Models are established
  for the organizations set of standard processes
  (OSSP)
• Process Performance Baselines Measurements of
  actual performance for the OSSP. Include a
  distribution of results that can be used by all
  projects for estimation purposes.
• Process Performance Models Predictors of future
  performance of the OSSP and of processes in a
  projects lifecycle may use simulations and
  other statistical methods
• A quantitative understanding of the OSSP enables
  individual projects to know how to compose/tailor
  their defined processes so that their objectives
  are realistic and can be met.

9
CMMI Maturity Level 4 Process Areas

• Quantitative Project Management (QPM) involves
  the following
• Establishing and maintaining the projects
  quality and process-performance objectives
• Identifying suitable subprocesses from the OSSP
  baselines and models to tailor and include at the
  project level
• Selecting the project subprocesses that should be
  statistically managed and monitoring their
  performance
• Sending appropriate statistical and quantitative
  data to the Organizational Measurement Repository
• Understanding the nature and extent of the
  variation experienced in the projects process
  performance
• Determining if the projects quantitative
  objectives for quality and process-performance
  can be met
• Ensuring that there is not a misalignment between
  the Voice of the Customer Specification Limits
  and the Voice of the Process Control Limits --
  desired results vs. capability
• The principles of QPM can be applied to projects,
  support groups, and functional areas

10
Quality Management System (QMS) Involving Value
Streams

• QMS methodology must support the initiative to be
  CMMI ML 3 compliant.
• Value Streams (VSs) can be used to select
  subprocesses for statistical management.
• Value Streams are representative of major
  lifecycle phases.
• SEI is recommending for QPM that at least one
  subprocess be statistically managed within each
  major project phase and that at least one
  subprocess be included from each of the four CMMI
  Continuous Representation Categories (Process
  Management, Project Management, Engineering, and
  Support)
• Statistically managed subprocesses should
  facilitate predicting process outcomes and
  tracking current project progress.
• Peer Reviews could be one of these control
  knobs/subprocesses to manage statistically as it
  can appear in several Categories.
• It may take two or more years to collect
  sufficient data to determine the critical
  subprocesses to optimize for ML 4.

11
Value Stream Structure

• Overview of process
• Identifies all associated sub processes
• Provides for CMMI artifact collection
• Includes other functionality (CM, ILS, etc.)

Installation Value Stream
CMMI ML 3 Process Areas
DAR
RSKM
REQM
CM
RD
TS
OT
IPM
PP
PMC
SAM
MA
PPQA
PI
VER
VAL
OPF
OPD
Pre-Install Phase
Install Phase
Closeout Phase
IDP
BESEP
Turnover
Site Survey
Install prep
Project initiation
Install Activities
Test Checkout
Sub processes
12
QA / CMMI Structure
Command Level Docs
RSKMP
CMP
SEP
MA P
RMP
SAM
QAP
ILSP
PMP
COI/PM Level Docs
COI/Program Management Plan
Project Level Docs
Project /Assignments Plan
Value Streams (handbooks)
Project Plans tie the COI to the VS
R D VS
Acquisition VS
Integration VS
Installation VS
Life Cycle
ISEA VS
SSA VS
Repair VS
13
Quality Management System (QMS) Involving Value
Streams

• The following Command processes complement and
  are embedded in the foregoing VSs
• Finance
• Contracting
• Purchasing
• Business Processes
• CMMI
• QA
• Etc.
• Note Logistics and Configuration Management may
  become VSs themselves if they are separate
  product/service deliverables to a customer in
  their own right.

14
LSS Focused Efforts

• Black Belts will assist the Implementation of
  CMMI
• Work with Divisions to determine VSs
• Break VSs down into Process steps
• Green Belts will conduct events to further define
  and improve VS Processes
• Black Belts will support the EntPG Structure
• Basic EntPG Communication Plan will provide
  visibility and coordination of these efforts

15
Detailed Communication Plan
EntPG 09K
Master BB
IPT Chair
IPT Chair
IPT Chair
IPT Chair
Dept EPG
Dept BB
DIPT Chair
DIPT Chair
DIPT Chair
DIPT Chair
Div BB
Div EPG
Div EPG
Div EPG
TWG Chair
TWG Chair
TWG Chair
TWG Chair
TWG Chair
TWG Chair
TWG Chair
TWG Chair
TWG Chair
TWG Chair
TWG Chair
TWG Chair
TWG Chair
TWG Chair
TWG Chair
16
8 Step Process

• The 8 Step Process is the improvement methodology
  based on tools and techniques of Lean Six Sigma

Identify Prioritize Opportunities
Project Definition
Document Measure Current Reality
Analyze Waste Variation
1
2
4
3
Excellence
Implement Validate
Improve Innovate
Communicate Acknowledge Success
Measure Sustain
5
6
8
7
17
Purpose of Measuring

• Measurements should do any or all of the
  following
• Directly tie in to the project objectives
• Assess current performance of your process
  against customer requirements
• Show baseline (existing) capability
• Validate project benefits and ensure visibility
  of improvements
• Identify relative strengths and weaknesses in and
  between processes
• Monitor and control long term performance
• Be concise and readily understandable
• Facilitate rapid understanding of what is
  happening in the process/project

18
Establishing a Measurement

• Establishing a measurement follows a generic
  process
• Five steps to establish an effective measurement
• Select the measurement
• Define the measurement
• Identify measurement/data source
• Define a measurement collection/sampling plan
• Maintain the measurement

19
Select the Measurement

• The selection of effective measurements should
  consider data that is
• Relevant
• Related to your purpose or goal and actually used
• Is collected real time
• Accurate
• Measures what was intended to be measured
• Obtained through careful reading and recording
• Understandable
• Clear and organized
• Summarized in tables or graphically
• Complete
• Includes every necessary detail and event
• Enables you to make decisions
• Simple
• Data is available without extraordinary effort
• Data is not expensive to collect

20
Measurement Effectiveness

• Need to define measures to assess project
  complexity. Possible candidates areas
• Estimated engineering hours (not dollars which
  could change) Technology risks
• Number of requirements and their volatility
• Management has a role in driving down complexity
  in all phases of a project.
• More consideration must be paid to ways of
  reducing Project Challenge Complexity. Doing
  so is a major challenge prior to the
  establishment of the development project,
  beginning during the pre-acquisition period.
  Earlier application of Systems Engineering
  practices and principles may go a long way
  towards reducing that challenge. Survey Report,
  p. 100

21
6s

• In statistics, s represents standard deviation, a
  measure of variation for process performance
• Sigma is generated by measuring processes
• Process data can be collected and evaluated to
  determine its impact on productivity,
  performance, and customer satisfaction
• The measurements provide the ability to predict
  process performance and provide a benchmark to
  determine if actions have produced results

22
Operating at 6? Capability implies

• Data driven decision making
• Meeting customers requirements
• Measurable processes
• Processes Under Control
• Variation has been reduced
• Future performance can be predicted
• Results of actions can be assessed

23
Utilizing the Data

• Process variation is large compared to
  specifications such that the process must remain
  centered to maintain capability
• Any shifts in process mean, or variation, will
  abruptly reduce the capability of the process
• Cp measures the ability of a distribution to fit
  within specs
• Cp (usl-lsl / 6s)
• Cpk measures how well the process is centered
• Cpk min (usl-m / 3s, m-lsl / 3s)

24
Utilizing the Data

• Cp and Cpk are used together so that they may
  indicate if the process can meet specs and is
  properly centered
• Cp
• Low variation
• Centered Cp
• Cpk High variation
• Not Centered
• Cpk
• Cp
• Low variation
• Not Centered
• Cpk

mean
target
USL
LSL
25
Interpreting Histograms

• Histograms can judge current performance and
  future performance

26
Wrap

• Six Sigma supports both individual process areas
  as well as aggregate process data
• Six Sigma statistical processes should be applied
  to rolled up management level aggregate data as
  well as singular process data
• Forecasting, modeling and prediction can be done
  utilizing statistically derived data
• Six Sigma provides a tool set that can be used to
  collect and analyze data used in high maturity
  organizations