Simple is Not Necessarily Better: Why Software Productivity Factors Can Lead to Bad Estimates
The use of simple linear productivity factors to estimate software development costs is rampant in DoD. Although these factors are easy to explain and use and are (usually) grounded on historical data, they nevertheless contain several cost estimating risks including:
• Ignoring the negative impact of size on effort (diseconomies of scale)
• Ignoring the impact of incremental development on effort
• Obscuring the system effect
• Encouraging the use of flawed adjustment factors to account for missing software development activities
This presentation discusses each of these risks and offers some suggested remedies to mitigate them. This presentation reflects our firsthand experience gained from current software cost estimating research and cost estimating support as well as the results from previous research efforts including a three-year Small Business Innovation Research (SBIR) project that investigated improved software development cost and schedule estimating techniques.
Mr. Gallo is a Senior Cost Analyst at Technomics, Inc. He has spent most of his 16 year career working with software intensive weapon systems, automated information systems and intelligence systems. In his current tenure at Technomics, Mr. Gallo has been conducting research into improved technqiues for estimating software development costs. Additionally, Mr. Gallo provides key consulting support to the Office of Secretary of Defense (OSD) Software Resource Data Reporting efforts. He is also the principal architect of VERA, Technomics’ software cost and schedule estimating tool.
Prior to joining Technomics in June 2000, Mr. Gallo was a senior cost analyst at TASC where he was responsible for managing, developing, and integrating life cycle cost estimates of software intensive systems in both DoD and the Intelligence community.
During his five year tenure at the Naval Center for Cost Analysis, Mr. Gallo was a contributing author to the creation of the Navy Software Estimating Handbook. Mr. Gallo was also responsible for developing independent life cycle cost estimates and economic analyses on various Navy and Marine Corps systems. He holds both a Master of Engineering degree and Bachelor of Science degree in Industrial Engineering from the State University of New York at Buffalo.
Mr. Hardin graduated from the Georgia Institute of Technology with a Bachelors degree in Industrial Engineering in May 1991. Upon graduation, he accepted a cost analyst position with the Naval Center for Cost Analysis (NCCA). While at NCCA, Mr. Hardin performed independent life cycle cost analyses on numerous Navy weapon system programs and led a joint US/UK effort in the development of the Operating and Support Cost Analysis Model (OSCAM). In April 1999, Mr. Hardin accepted a position with the Office of the Secretary of Defense (OSD), Cost Analysis Improvement Group (CAIG). After joining OSD, Mr. Hardin received his Master’s degree in Systems Engineering from the Virginia Polytechnic and State Institute in May 1999. In January 2000, Mr. Hardin left the government to take a financial analyst position with Broadband Office. While at Broadband, Mr. Hardin performed financial analyses and developed models in support of Data and “Voice over IP” infrastructure designs. In January 2001, Mr. Hardin accepted a cost analyst position with Technomics, Inc. Since joining Technomics, Mr.
Hardin has developed improved methodologies for estimating the cost of manned and unmanned ground combat vehicles, radar systems, Commercial Off-the-Shelf (COTS) hardware impacts, ship engineering and production, software development, and system integration and has performed cost analysis for the DDX, LCS, CGX, CVN, USCG Deepwater and VAMOSC programs.