Download Capability Maturity Model

Capability
Maturity
Model
Part One - Overview
History

1986 - Effort started by SEI and MITRE
Corporation

assess capability of DoD contractors

First version published in 1991

closely related to TQM

goal is customer satisfaction

not required that customer be "delighted"
Some Fundamental Ideas

Process improvement is based on
small steps, rather than revolutionary
innovation.

CMM is not exhaustive or dictatorial.

CMM focuses on processes that are of
value across the organization.
Benefits of using the model

helps forge a shared vision of purpose of
process improvement within the organization

establishes common language for the software
process

defines a set of priorities for attacking problems

supports measurement
 via reliable appraisals
 objective
 supports industry-wide
comparisons
Risks of using the model

"All models are wrong; some models are
useful."

Not a silver bullet.

Does not address all of the issues that are
important for successful projects.

For example
how to hire and retain good people
 expertise in the application domain

Levels
1.
2.
3.
4.
5.
Initial
Repeatable
Defined
Managed
Optimizing
Level 1 : The Initial Level

ad hoc, sometimes chaotic

overcommitment leads to a series of crises

during a crisis, projects abandon plans

capability is characteristic of individuals,
not the organization

when a good manager leaves, the success leaves
with them
Level 2 : The Repeatable Level

Planning is based on experience with
similar projects


past successes can be repeated
Policies for Managing and Implementation

installed basic management controls
 track costs and schedules
 notice and deal with problems as they arise
Level 3 : The Defined Level

Standard Processes defined across the
organization and used by all projects
standard set of roles, activities, quality tracking, etc
 each project uses a tailored version of this
standard process


Training Program is in place to ensure
everyone has the skills required for their
assigned role
Level 4 : The Managed Level

Quantitative Quality Goals


Organization-wide Process Database

meaningful variations in process performance
can be distinguished from random noise


for both Products and Processes
actions are then taken to correct the situation
Products are of predictably high quality
Level 5 : The Optimizing Level

Organization has the means to identify
weaknesses and strengthen the process
proactively

teams analyze defects to determine their cause,
and disseminate lessons learned throughout the
organization

major focus on eliminating waste

e.g. reduce amount of rework
Defect prevention
Technology change management
Process change management
Key Process Areas
by maturity level Quantitative process management
Software Quality Management
Organization process focus
Organization process definition
Training program
Integrated software management
Software product engineering
Intergroup coordination
Peer Reviews
Requirements management
Software project planning
Software project tracking and oversight
Software subcontract management
Software quality assurance
Software Configuration management
This is a
somewhat handy
hierarchy of activities.
Don't skip levels

For example,

collecting detailed data (level 4) is
meaningless unless the data is from
projects that use a consistent process
(level 3)
Reality Check...

Is CMM well-suited for
everyone?
Criticisms of CMM

CMM is well suited for bureaucratic organizations
such as government agencies and large
corporations.


Its formality is a hindrance to projects where time-tomarket is more important than quality.
Promotes process over substance.

For example, emphasizing predictability over service
provided to end users.
en.wikipedia.com
Who uses CMM

75% of organizations are probably
Level One.


To get to Level Two, you must have control
over the requirements documents. Hence,
shrink-wrap companies like Microsoft are
Level One.
75% of Level Five organizations are in
India.
Level Comparison - Risk

Level 1
 Just

do it
Level 2
 problems

are recognized and corrected as they occur
Level 3
 problems
are anticipated and prevented, or impacts
minimized

Levels 4 and 5
 sources
of problems are understood and eliminated
Level Comparison - People

Level 1



Level 2




people are trained for their role(s)
groups work together
Levels 4


people are trained and supported by management
success depends on individuals
Level 3


success depends on individual heroics
fire fighting is the way of life
strong sense of teamwork in every project
Level 5


strong sense of teamwork across the organization
everyone does process improvement
Level Comparison - Measurement

Level 1


Level 2



data collected for all processes
data shared across projects
Levels 4


individual projects use planning data
Level 3


ad hoc data collection and analysis
data standardized across the organization
Level 5

data used for process improvement
Quiz 1
Your Role : SQA specialist
 Situation :


Initial Unit Testing reports indicate a bug rate
of 4.5 / KSLOC.
 Further checking finds
Average initial bug rate is 3.1 per KSLOC
 StdDev of 0.5
 weighted rate is also higher than average


What CMM level enables this amount of
visibility into the process?
Quiz 2


Your Role : Project Manager
Phase : Unit Testing

Problem : You notice that design,
implementation, and testing of the database
component will probably take 3 weeks instead of
the planned 4 weeks. These tasks are not on
the critical path.

Possible Actions?
 and
what level CMM characterizes that action?
Quiz 3

Your Role : Project Manager
Phase : Planning

Task : Schedule the Testing of the Database

 Estimated
Duration: 3 days
 Required Resources:





the database requirements specs
the implementation (source code)
real data from customer
test person that has a DB Test certificate
How do the different capability levels affect
your ability to schedule and monitor this
task?
Quiz 4


Your Role : a development team leader
Problems :
 Well
into development, you get an email indicating a
change in the interface requirements based on a demo
of the prototype done with the customer. The change
will require a good amount of code rework.

What would be the reaction of groups with
 Level
One
 Level Three
 Level Five
Next . . .
 Key
Process Areas
 Key Practices
 Common Features