Estimating intrinsic engineering value of emerging software innovations

Visveshwar (Vishu) Reddy, Pace University

Abstract

The software industry is characterized by a significant number of mergers and acquisitions and licensing transactions each year. When the driving factor for these transactions is technology acquisition, the effort to recreate the innovation by the potential licensee becomes a key variable to pricing and negotiations. There are algorithmic models that have been in existence for nearly 40 years to estimate the engineering effort. However, due to calibration, accuracy, and other challenges associated with algorithmic models, non-algorithmic models (e.g., expert judgment, analogy) are most widely used in the industry despite challenges such as repeatability and scalability. Neither the algorithmic nor non-algorithmic estimating studies comprehensively address what it would take for a potential licensee to recreate an emerging software innovation. The focus of this research is to provide a framework for calculating a metric called Intrinsic Engineering Value, or IEV, which represents the person years of effort required to reproduce the same emerging software innovation functionality. The study captures the problem in terms of qualitative attributes and uses conjoint analysis and wisdom of the crowds to estimate the IEV of emerging software innovations. The research was conducted in fives phases in the field at the IBM T.J. Watson Research Center in Hawthorne, NY. Phase 1 identifies preliminary IEV attributes for further analysis. Phases 2 and 3 establish attributes levels and use orthogonal experimental design to develop 18 hypothetical product combinations. Phase 4 collects 18 IEV estimates from each of the 50 IBM researchers involved with developing emerging software innovations. Phase 5 analyzes these 900 IEV estimates using ordinary least squares regression, determines the part-worth of each attribute, and establishes a quantitative model to estimate the effort required by a potential licensee to recreate an innovation. In the 11 case studies across eight different domains used for validation, the model estimate was within the range or close to the expert estimate in most cases. Multiple applications for the model are outlined. The research can be extended for general use to quantify qualitative problems and develop parameterized models.

Subject Area

Management|Computer science

Recommended Citation

Reddy, Visveshwar (Vishu), "Estimating intrinsic engineering value of emerging software innovations" (2010). ETD Collection for Pace University. AAI3401495.
https://digitalcommons.pace.edu/dissertations/AAI3401495

COinS