Data-Driven Estimating – Quantifying Electronics Complexity Using Public Data
Methods and Models II Track
The parametric cost estimation industry is being challenged to increase the defendability of their estimates, mainly through comparison to analogous data. However, companies are more protective than ever of their high quality cost data. While this data is often obtainable, it usually comes with strings attached (non-disclosure agreements, etc.) that preclude its usage in defending your estimate. At the same time, government and military databases with cost and technical data are being exposed to the public, and while the noise introduced by granularity issues and the mapping process of this data presents its own set of challenges, this trend also contains the key to a solution.
This paper discusses our research into quantifying electronics complexity for inclusion in our hardware cost estimation model, while meeting the challenges of this new reality. I will discuss our research approach to this problem, our final solution structure, and interesting insights obtained along the way. The process is based on data obtained with the IHS Haystack tool, which contains parts and logistics data on over 100 million items in the U.S. Federal Supply Catalog across 70+ Army, Navy, Air Force, and other government, military and commercial databases. I will discuss issues we encountered in molding available data to a format that fits our cost models, validating our assumptions, and dealing with incomplete data. The end result is a calculator that guides users through quantifying complexity of a vast selection of electronic items in a way that is defendable by comparison to the public data on which it is based.
F. Gurney Thompson III
F. Gurney Thompson III is a cost research analyst at PRICE Systems LLC. With progressive experience in mathematical modeling and a strong conceptual mind, Gurney is ideally suited for research and development of cost methodologies and analysis. Most recently, he has lead a major effort in enhancing tools for electronics cost estimation, while also developing approaches for increasing defendability of parametric cost estimates. In 2010-2011, Gurney was the primary architect of solutions for estimating total ownership cost at PRICE Systems. Gurney was a major contributor to the US Army Software Lifecycle Affordability Management study in 2009-2010, where he analyzed cost implications of Service Oriented Architecture. Gurney graduated with a B.S. in Systems Engineering from the University of Virginia in 2007.