CMMI High Maturity Best Practices HMBP 2010: Deploying High Maturity Practices Globally by Kaliappan M

©2010 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice1

Deploying High Maturity Practices Globally

Kaliappan M, Head Quality - HP

©2010 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice2

Kaliappan, M

May 2010

Hewlett-Packard

Deploying High Maturity

Practices Globally

3 Deploying High Maturity Practices Globally

Overview

1. Challenges to High Maturity

2. Intent of HM Practices

3. Foundational elements

4. Deploying HM Globally

5. Recap and Q&A

– While High Maturity deployment

in general has many

challenges, implementing them

across global centers pose

unique challenges.

– To gain maximum business

benefit, the intent of the HM

practices need to be understood

and the critical aspects of

deployment to be considered.

– This presentation will address

the challenges as well as

essential elements for robust

deployment based on the

experience.


Section 1 – Challenges to High Maturity





5. Recap and Q&A1


High Maturity Challenges

– To maximize business benefits, the

HM implementation needs to be

effective

– Greater attention to detail in

deployment is expected during

HM Appraisals

– Increased competition in the

market place

– Lead time to reach higher maturity has

come down from CMM days

– Global deployment challengesCharts from SEI Maturity Profile Aug 92 and Mar

10 Respectively


Major Obstacles for High Maturity

– Lack of contextual information for proper segmentation/stratification

– Not enough consistency in measures to aggregate/disaggregate them

– Resistance to collect new or additional measure after achieving ML 3

– Didn't collect data frequently enough to make mid-course corrections

– Not enough coaching to the performance model developers

– Process Performance Model developers didn't have enough statistical

expertise or access to experts

– Spent too much time in preparing Management Review report rather than

doing thorough analysis

– Models focus only on final project outcomes rather than

interim outcomes

– Failed to align and prioritize M&A practices to the business objectives

Note: Listed from high to Low order Ref: SEI Technical Report 2008-TR-024


Section 2 – Intent of HM Practices





5. Recap and Q&A2


Baselines and Models

Controllable and

segmenting factors

Sub-process measures

and indicators

Basic measures

– Do not generalize

across domains and

project types

– Do not aggregate data

from heterogeneous

processes

• Homogeneity of data

leads to meaningful PPB

analysis

• Homogeneity is ensured

by segmentation and

stratification

– Ensure Operational

Definitions are available

at adequate level of

detail


Characterizing the Process/Sub-process

Process performance

baselines/ Models

Cause- effect relationships

Context (domain and tailoring

impact)

Statistics (dispersion and central tendency)

and distributions

Stability and capability

information

Inter-relationships of measures, facto

rs, actions


Ingredients of Performance Baselines

Addresses measures from processes and sub-processes

Distribution is known

Appropriate central tendency and dispersion statistics used

Ensure homogeneity in data while grouping

Use Statistical techniques to decide significance and

segmentation

Indicates the stability and capability of the sub-processes

Demonstrates trends between baselines

Includes data on controllable factors (‘x’) as well as responses

(‘y’)

Supports the projects in 'composing' the process


Controllable Factors

– Has influence over one or more project outcomes

– Controllable factors relates to any of the following:

• People, Environment, Technology, Tools, Process and Stakeholders

Controllable Factors (X) Response (Y)

− Review types

− Design method

− Testing type

− Language

− Design and code complexity

− Requirements Volatility

− Test coverage

− Review coverage

Delivered defect density

Controllable Factors (X) Response (Y)

− Domain experience(H, M, L)

− DLC Model

− Complexity

− Staff turn over

Productivity

− Estimation method

− Estimator skill

− Project type

− Complexity

− Staff turn over

Cost and Effort Variance


Influence of Controllable Factors

Design Review

Design

Review Method

Review Efficiency

Preparation rate

Reviewer Skill

Design Defect Density

TestingProduct

Test Cases

Testing Type

Test Coverage

Code Coverage

Tester Skill

Test Defect Density

Code Review

Code

Review Method

Review Rate

Preparation rate

Reviewer Skill

Code Defect Density

Review Session Duration

Size/ Review session

Delivered Defect Density

Leading Indicators

LaggingIndicators

PPM1

PPM2

PPM3

PPM4

This is defect detection example. Similarly, defect prevention factors can be related.


Controllable Factors in PPMs

Code Review

Code

Review Method

Review Rate

Preparation rate

Reviewer Skill

Code Defect Density

Review Session Duration

Size/ Review session

3.02.52.01.51.00.50.0

12

10

8

6

4

2

0

Code Review Rate(Person-hours/K

Co

de

Re

vie

w D

efe

ct

De

nsit

y(D

efe

S 0.848127

R-Sq 94.3%

R-Sq(adj) 93.2%

Fitted Line PlotCode Review Defect Density(Defe = 0.8667 - 2.928 Code Review Rate(Person-hours/K

+ 2.627 Code Review Rate(Person-hours/K**2


Ingredients of Performance Models

Statistical, Probabilistic or Simulation in nature

Logical relationship between X's and Y

PPMs developed / calibrated based on ORG data

Predicts interim or final project outcomes

Connects 'upstream' activity with 'downstream' activity

Prediction & Confidence levels supported

1 or more Controllable & 0 or more non-controllable factors

Factors relate to sub-process to conduct prediction

Enable projects to achieve mid-course corrections

Enable 'What-if Analysis' for planning, dynamic re-planning

and problem resolution

Extracted from SEI presentation ‘A tutorial on CMMI Process Performance Models’


Sub-process Management

– Statistical management at sub-

processes than overall project level

• Measurements at project level are

masked by behavior of multiple sub-

processes

• Focus on ‘Analysis measures’

– Use statistical methods to analyze,

understand and characterize sub-

process variation

• Control charts, Box and Whisker plots,

Histograms, etc

– Do not simply implement control charts

• Understand the special cause patterns

• Implement in real time

• Investigate for multiple cause systems


Analyzing for Root Causes

– Not just defects and problems

• Achieving or non-achieving of

objectives; and improving

performance

– Look for both types of causes

• Causes of successes and failures

– Reactive as well as proactive

– Sound outcome selection for CAR

analysis

• Using Pareto, Histograms, Box and

Whisker plots, capability analysis

– Judge improvements for statistical

significance when applicable


Composing a process

Composing a process

involves a quantitative

analysis of one or more

alternatives for the project’s

defined process to determine

the probability of achieving

the project’s quality and

process performance

objectives

Delivered Defect Density

Design Review

Design Review

Checklist

Review Method

Preparation rate

Reviewer Skill

Design Defect Density

TestingProduct

Test Cases

Testing Type

Test Coverage

Code Coverage

Tester Skill

Test Defect Density

Code Review

Code Reuse

Review Method

Review and Preparation rate

Reviewer Skill

Code Defect Density

Release

Rework

Coding

Code Review Checklist

Design

RA Method

Complexity

Reqts. Review

RA

Design Method

Reqts. Defect Density

Review Method

Reviewer Skill


Evolution of Process Composition

Level 2− Select right LC

− Identify phases/

processes based

on scope

Level 3− Use past project

learning

− Fine tune

processes/method

s qualitatively

through tailoring

Level 4− Use PPBs for sub-

process selection

− Use PPMs with

controllable factors

− Fine tune sub-

processes/methods

quantitatively

Level 5− Simulations used to

link sub-process

capabilities to

project goals

− Sub-process/

methods selection

based on trade-off

decisions

Still linking sub-process capabilities to goal achievement might be a challenge


Section 3 – Foundational Elements





5. Recap and Q&A3


Constancy of Purpose

– Focus all actions on one aspect –

meeting business objectives

– Demonstrate High maturity by

building business value

– Target both process as well as

product/service quality

improvements

– Maximize structured decisions

based on data

– Eliminate waste – do not over do

• Decision processes, Causal

analysis, Statistical analysis, Sub-process

management and model development


Systems Thinking

Processes are Interconnected

– Multiple sub-processes need to be monitored to validate

achievement of objectives

– Multiple measures and factors needed to understand a sub-

process

Never attempt implementing HM PAs in isolation

– OPP, QPM, OID and CAR – all are highly inter-related

– HM PAs interact with lower level PAs


HM Process Interactions

QPPO – Quality and Process Performance Objectives

QPPO

QPPOOU level

OPP

OID

DAR

CAR

QPPO

Project Level

QPPO

Sub-Process

Objectives

QPM

CAR

QPM

CAR

‘Quantitative Control’

‘Statistical Control’

PPM

PPB

Set of sub-processes

OU level

PPM

PPB

Set of sub-processes

MAMA

IPM

RSKM

RSKM

OPD

VER

VAL

PMC

PP

MADAR

RD, TS


Robust Measurement System

– Automation is critical for HM

Initiative’s success

• Enhances data quality aspects;

• Facilitates timely analysis and actions

– Inferential analysis is critical in High

Maturity journey

• Analysis done more often;

• Inferential analysis as against status

reporting

– Facilitates real time analysis and action

• Data collection, data analysis, Root

cause analysis and taking corrective

action in real time


Robust Measurement System (cont’d)

– Granularity and quality of data

• Granularity of data collection needs to be

at the right level

• Increased usage levels of

progress, success and analysis

indicators

• Quality of data collected

– Optimize cost of data collection

and analysis

• More granularity only for those sub-

processes selected

• Sub-processes and measures aligned to

the objective – instrument them as

needed


Enablers for Initiative’s Success − Survey

0

2

4

6

8

10

12

14

# R

esp

on

de

nts

Enablers for Delivery Excellence

Based on a short survey conducted by the author across IT industry in March 2010.


Infrastructure

Constancy of Purpose

Integration

Rewards and Recognition

Iterative Approach

• Core team

• Virtual team

• Goal Deployment

• Metrics program alignment

• Integrated Initiatives

• Business Integration

• Positive Reinforcement

• Iterate for Quick Wins

Foundation for Initiative’s Success


Critical Elements of Initiative Process

Implement

Celebrate &

LearnPrepare

Strategy

Develop

Capability

Readiness

Assessmen

t

Initiative

Design

Impact

Analysis

Detailed

Planning

Monitoring

Communication

Education & Training

Stakeholder Involvement


Section 4 – Deploying HM Globally





5. Recap and Q&A4


HM Capability Development

HM Capability

Development

Strategy

Models and

Baselines

Deployment

ManagementAssets

HM

Communit

y

Capability

Assessme

nt

− Interim Checks

− Coach Qualification

− Best Practice Sharing

− Enable reuse

− Training

− Guidance

− Planning

− Execution

− Development

− Reviews


Training

Team

Members

Project

Managers

Model Developers

& Coaches

Team

MembersQuality

Managers

− Introduction to HM Concepts

− High Maturity Workshop for Managers

− Advanced Metrics and Statistics

− Introduction to HM Concepts

− HM Deployment Management


HM Roles

HM

Consultant

Model Developers

and HM Coaches

Project Managers

Team Members

− Having experience in HM implementation/consulting

− Either groomed within the Organization/or in the region

− Completed ‘Advanced HM Training’ and passed

an ‘Exam’

− Successfully completed ‘High Maturity

Workshop for Managers’

− Introductory courses in High Maturity

Concepts


– Effectiveness of High Maturity

Deployment is impacted by various

factors within

the Organization

• Organization Context

• Process Improvement history

• Foundations for ML2 and ML3 for High Maturity

• Modeling and Baseline considerations

• Resourcing considerations

– HM Initiative to be managed as a project

in an iterative life cycle

HM Deployment Management


Approach to High Maturity

– HM Planning

– HM Initiative Duration Estimation

– Training & Qualification

– Initial Gap Analysis

– Create PPMs and Baselines

– Deploy processes

– HM Coaching

– HM Interim Checks

• HM Risk Assessment Grid

– SCAMPI-B Appraisal

– PIIDs

– SCAMPI –A

Appraisal

HM Initiative Duration

Plan Execute Appraise


HM Risk Assessment Grid

Target Grid Mapping

HM Code Check item Attributes

Grid Attributes

Risk Reporting


Section 5 – Recap and Q&A





5. Recap and Q&A5


Recap

Understanding the

challenges and intent of

High Maturity practices helps

structure our actions for

deployment effectively.

Adequate focus on the

foundational elements help

not only enhancing the

business benefits, but also

accelerate the HM

deployment.


Questions?

M. Kaliappan

HP Best Shore Application Services

Hewlett-Packard

HP Avenue, 39/40 Electronic City,

Bangalore – 560100

+91 80 26182123

[email protected]

mailto:[email protected]


Outcomes that matter.


Click here for:

High Maturity best practices

HMBP 2010 Presentations

organized by QAI

Click here

http://www.qaiglobal.com/qgs/HMBPListing.asp?ID=800



http://www.qaiglobal.com/
