Measuring Dollar Savings from Software Process Improvement with COCOMO II Betsy Clark Software Metrics Inc. October 25, 2001 Acknowledgment: This presentation.

Slides:



Advertisements
Similar presentations
Paul J. Frenz General Dynamics Advanced Information Systems
Advertisements

Cost as a Business Driver 1 John Brown C Eng MIEE mr_ Software Cost Estimation.
McGraw-Hill/Irwin Copyright © 2007 by The McGraw-Hill Companies, Inc. All rights reserved. 5 Capacity Planning For Products and Services.
1 State of Michigan Achieving Software Process Improvement with Capability Maturity Model (CMM)
Management: Analysis and Decision Making
Software Process Improvement Robin B. Hunter, Ph.D. Vol 2., p Presented by: Andrew Wheeler.
CPIS 357 Software Quality & Testing I.Rehab Bahaaddin Ashary Faculty of Computing and Information Technology Information Systems Department Fall 2010.
SAFETY & QUALITY MANAGEMENT IMPROVE SAFETY WITH QUALITY & PRODUCTIVITY BENEFITS Dave Copley, Matt Pierce.
© Copyright Richard W. Selby and Northrop Grumman Corporation. All rights reserved. 0 Process Synchronization and Stabilization February 2007 Rick.
A framework for describing IT Project Management Processes and Tool Set Features Enterprise Project Management Framework.
Software Quality Processes – Part I CSSE 376, Software Quality Assurance Rose-Hulman Institute of Technology March 16, 2007.
Software Quality Engineering Roadmap
Capability Maturity Model (CMM) in SW design
SE 450 Software Processes & Product Metrics Software Metrics Overview.
University of Southern California Center for Systems and Software Engineering Assessing the IDPD Factor: Quality Management Platform Project Thomas Tan.
Glen Knight, PMP, CSP President How Mature Do You Think Your Are? The Project Management Maturity Model.
Readiness Index – Is your application ready for Production? Jeff Tatelman SQuAD October 2008.
Software Engineering Institute Capability Maturity Model (CMM)
Capability Maturity Model
Cost Management Week 6-7 Learning Objectives
1. 2 What is Six Sigma? What: Data driven method of identifying and resolving variations in processes. How: Driven by close understanding of customer.
Managing Project Quality
Six Sigma By: Tim Bauman April 2, Overview What is Six Sigma? Key Concepts Methodologies Roles Examples of Six Sigma Benefits Criticisms.
© The McGraw-Hill Companies, Software Project Management 4th Edition Software effort estimation Chapter 5.
9 - 1 ©2002 Prentice Hall Business Publishing, Introduction to Management Accounting 12/e, Horngren/Sundem/Stratton Chapter 9 Management Control Systems.
TTMG 5103 Module Techniques and Tools for problem diagnosis and improvement prior to commercialization Shiva Biradar TIM Program, Carleton University.
© Mahindra Satyam 2009 Project Metrics QMS Training.
Copyright © 2013 by The National Restaurant Association Educational Foundation. Published by Pearson. All rights reserved. HOSPITALITY HUMAN RESOURCES.
N By: Md Rezaul Huda Reza n
Cost of Quality - COQ MGMT-5060 Operations Management.
PPMT CE-408T Engr. Faisal ur Rehman CED N-W.F.P UET P.
This document is proprietary to Project Consulting Group, Inc. and contains confidential information which is solely the property of Project Consulting.
Software Estimation and Function Point Analysis Presented by Craig Myers MBA 731 November 12, 2007.
Software Engineering Lecture # 17
University of Southern California Center for Software Engineering C S E USC Using COCOMO for Software Decisions - from COCOMO II Book, Section 2.6 Barry.
“Look, who is the most successful in attracting and holding good people? The nonprofits. The satisfaction has to be greater than in business because there.
Software Product Line Material based on slides and chapter by Linda M. Northrop, SEI.
Kathy Corbiere Service Delivery and Performance Commission
The Implementation of BPR Pertemuan 9 Matakuliah: M0734-Business Process Reenginering Tahun: 2010.
What Have we Learned: Return-on- Investment from the SW-CMM Khaled El Emam v
Chapter 7: Project Cost Management
SOFTWARE PROCESS IMPROVEMENT SHARATH CHANDAR REDDY ALETI CSC 532 TERM PAPER.
What is project management?
SOFTWARE PROCESS IMPROVEMENT
Info-Tech Research Group1 Manage IT Budgets & Cost World Class Operations - Impact Workshop.
Project Planning Goal 1 - Estimates are documented for use in tracking and planning project. Goal 2 - Project Activities and commitments planned and documented.
Done By: Asila AL-harthi Fatma AL-shehhi Fakhriya AL-Omieri Safaa AL-Mahroqi.
Capability Maturity Model. CS460 - Senior Design Project I (AY2004)2 Immature Organisations Software processes are often rigorously followed. Organisation.
The Marketing Plan Chapter 2. Section 2.1: Marketing Planning  Good marketing requires good planning Research your company Study your business environment.
1 P-CMM ® INITIATIVE Overview. 2 B A D C Improving Organisation Organisational Capability Technology ProcessPeople.
1 These slides are designed to accompany Software Engineering: A Practitioner’s Approach, 7/e (McGraw-Hill 2009). Slides copyright 2009 by Roger Pressman.
9 - 1 Chapter 9 Management Control Systems and Responsibility Accounting.
Chapter 5: Software effort estimation
1 Using COCOMO for Software Decisions - from COCOMO II Book, Section 2.6, 6.5 LiGuo Huang Computer Science and Engineering Southern Methodist University.
Capability Maturity Model. What is CMM? n CMM: Capability Maturity Model n Developed by the Software Engineering Institute of the Carnegie Mellon University.
TANGUY BERTOLUS Technical Director.
CS 577b: Software Engineering II
Project Cost Management
State of Michigan Achieving Software Process Improvement with
Verifying – Evaluating Software Estimates
Using COCOMO for Software Decisions - from COCOMO II Book, Section 2
Using COCOMO for Software Decisions - from COCOMO II Book, Section 2
Chapter 5: Software effort estimation
Capacity Planning For Products and Services
Capacity Planning For Products and Services
Capability Maturity Model
Capability Maturity Model
Using COCOMO for Software Decisions - from COCOMO II Book, Section 2.6
Using COCOMO for Software Decisions - from COCOMO II Book, Section 2
Capacity Planning For Products and Services
Presentation transcript:

Measuring Dollar Savings from Software Process Improvement with COCOMO II Betsy Clark Software Metrics Inc. October 25, 2001 Acknowledgment: This presentation describes work being done by TeraQuest Metrics

Outline Background Measuring the Impact of Software Process Improvement (SPI) Some Initial Results

Customer Background Large financial institution Actively involved in software process improvement (SPI) –Software-CMM –System Test Began summer of 2000 at CMM Level 1 Incrementally adding Key Process Areas Two pilot organizations –Planning Level 2 assessment end of this year

Background (continued) Strong emphasis on measuring impact of SPI, especially hard dollar savings CIO: “If process improvement saves us money, I should be able to go down the street to my competitor’s bank and get a loan to fund our process improvement initiative.”

Outline Background Measuring the Impact of Software Process Improvement (SPI) Some Initial Results Conclusions

“Maturity levels are meaningless if they cannot be explained in terms of business objectives” John D. Vu Boeing Level 5 Organization

Business Objectives Reduce the cost of software activities Reduce delivery time Improve product quality Increase customer satisfaction –customers are internal to the bank (e.g., wholesale and retail mortgage, investment division)

Measurement Objectives Measure impact of SPI in terms of these business objectives Impacts of SPI will be measured by comparing a set of baseline projects to pilot projects

Measuring Hard Savings CFO’s initial understanding - –“If we have savings from SPI, we can reduce IT budget in the future.” –First point of discussion - need to measure work load –Led to concept of unit savings, holding IT organization accountable for those savings Brought IT manager into the discussion - –“But events occur outside of my control that can affect unit costs. For example, I can lose my top staff.”

Measuring Hard Savings The IT manager was talking about variability due to factors outside of SPI. That variability is addressed by parametric cost models. Approach - measure COCOMO II cost drivers for baseline projects and for SPI projects. Use them to adjust unit costs. –Backout all influences on unit costs except SPI

Measuring Hard Savings (cont) Savings due to SPI –Difference in adjusted unit costs between baseline and SPI projects

Setting Expectations SPI is a staged, long term initiative –implemented on pilot projects first, then on a wider scale Initially, we will estimate savings based on pilot results –few data points, wide variation As SPI is implemented on a wider scale, we will have more data points, clearer trends Moving from CMM Level 1 to Level 2 lays the foundation for unit cost savings –a few studies do show cost savings from Level 1 to 2 major effect is in better estimation and planning reduction in rework due to stable requirements

Measures 1) estimation accuracy: effort 2) estimation accuracy: schedule 3) productivity 4) unit costs 5) project delivery rate (cycle time) 6) system test effectiveness 7) delivered defect density 8) customer satisfaction 9) requirements volatility

Approach Attempted to “mine” existing data sources (e.g., time tracking, financial, problem reporting systems) –not successful, sporadic and inconsistently used Selected a representative set of completed projects from the two pilot organizations Goal was projects per pilot organization –13 projects from one –11 from the other Constructed a survey, met with project managers to collect data Followed-up with each manager to verify data

Measures 1) estimation accuracy: effort 2) estimation accuracy: schedule 3) productivity 4) unit costs 5) project delivery rate (cycle time) 6) system test effectiveness 7) delivered defect density 8) customer satisfaction 9) requirements volatility

Planned Labor Hours Percent difference between actual and estimated 0 Overruns Underruns Calculation: (Actual labor hours - estimated) / estimated Estimation Accuracy - Effort

Planned duration Percent Difference between actual and estimated 0 Overruns Underruns Calculation: (Actual calendar months - estimated) / estimated Estimation Accuracy - Schedule

Measures of Interest Median - very stable across organizations standard deviation Goals with SPI: –median should approach zero –standard deviation should be smaller

Measures 1) estimation accuracy: effort 2) estimation accuracy: schedule 3) productivity 4) unit costs 5) project delivery rate (cycle time) 6) system test effectiveness 7) delivered defect density 8) customer satisfaction 9) requirements volatility

Productivity and Unit Costs High variability Median is stable across divisions

Initial Results Used COCOMO II parameters to adjust size Led to a reduction in the standard deviation Helped explain: –why lower productivity projects had difficulty –why higher productivity projects had an easier time Projects with very high productivity seemed to do everything right –capable staff, low turnover, managing requirements… –these are good things that should improve with SPI –don’t want to penalize organization for improvement in these other (non-SPI) areas –management controllables vs noncontrollables

Measures 1) estimation accuracy: effort 2) estimation accuracy: schedule 3) productivity 4) unit costs 5) project delivery rate (cycle time) 6) system test effectiveness 7) delivered defect density 8) customer satisfaction 9) requirements volatility

Project Delivery Rate Calculation: –Function points / calendar months Goal: Increasing

Function Points Function points per calendar months Project Delivery Rate

Measures 1) estimation accuracy: effort 2) estimation accuracy: schedule 3) productivity 4) unit costs 5) project delivery rate (cycle time) 6) system test effectiveness 7) delivered defect density 8) customer satisfaction 9) requirements volatility

System Test Effectiveness Calculation: –(Defects Found in System Test / Total Defects) – where – Total Defects = (Defects Found in System Test + Delivered Defects found in first 30 days) –Example: –Defects found in System Test = 45 –Defects found in first 30 days of operations = 5 –Test Effectiveness = 90% Goal: 100% Result: Wide variation in effectiveness

Measures 1) estimation accuracy: effort 2) estimation accuracy: schedule 3) productivity 4) unit costs 5) project delivery rate (cycle time) 6) system test effectiveness 7) delivered defect density 8) customer satisfaction 9) requirements volatility

Delivered Defect Density Calculation: –Defects found in first 30 days of operations / function points Goal: 0

Function Points Defects per function points 0 COTS Custom Delivered Defect Density

(Very Preliminary) Finding of Interest In contrast to custom development, defect density for COTS projects appears unrelated to size

Measures 1) estimation accuracy: effort 2) estimation accuracy: schedule 3) productivity 4) unit costs 5) project delivery rate (cycle time) 6) system test effectiveness 7) delivered defect density 8) customer satisfaction 9) requirements volatility

Customer Satisfaction, Rqts Volatility Data do not exist Strategy was altered to request the manager’s estimate

Message to Executive Level Measurement –can be a powerful foundation for understanding and managing IT –is a cultural change and not a scoreboard –will improve as process maturity improves

Response from Executive Level (CIO and direct reports) Intense interest in the measures and in benchmarking Basis for excellent discussions about need for visibility into –requirements management –quality –customer satisfaction Collection of the nine measures has been made part of executive compensation –Moving forward to put supporting processes, tools and training in place

To be continued...