Case Study
SEER for Software and CMMI Equal High Maturity Software Cost Estimation
Dr. Richard Welch - Technical Fellow, Northrop Grumman Integrated Systems
Well-publicized industry data paint a dismal picture about the practice of software cost estimation and the accuracy of software cost estimates. Published statistics and post mortems typically point to poor initial cost estimation as a principle contributor to program risk, leading to cost and schedule overruns during project execution. Government cost evaluators are tasked to scrutinize cost proposals heavily and to suspect overly optimistic and poorly justified contractor claims.
The Northrop Grumman Integrated Systems facility in Melbourne, Florida, is a prime contractor for large software-intensive systems. We have a 15+ year track record of accurate software bids and successful project execution that have resulted in median cost under runs of 2%. How do we do it?
We have strongly embraced the concept of parametric cost estimation using SEER-SEM, and embedded its practice in the context of highly mature (CMMI Level 5) management and engineering processes. We use SEER-SEM to generate line-of-code-driven estimates that satisfy key Government risk items and demonstrate program executability.
CMMI goals and practices shape certain estimating behaviors, resulting in an integrated set of high maturity practices that have been incorporated into our software estimation process. This strategy incorporates lessons learned from successful large system proposal efforts and multiple non-advocate reviews and independent cost evaluations.
