Value of information in portfolio decision analysis

It can be time consuming to use decision analysis to allocate resources over a portfolio of projects. It may be possible to attain most of the value added by decision analysis in significantly less time. This paper defines and compares different analytic strategies in terms of the resulting value added for a range of simulated portfolios. A portfolio consists of a set of candidate projects or investments of uncertain value. The value of each project may be estimated with or without the additional information provided by decision-analytic methods. Projects are then completely prioritized and funded in order of the ratio of their expected net present value to their cost, or partially prioritized and funded whenever this ratio exceeds a predefined threshold level. The portfolio funded under the various analytic strategies is compared against a strawman alternative of random funding decisions. The value added through disciplined prioritization often exceeds the additional value added through the more costly step of developing refined estimates of project values. Intermediate approaches, including threshold approaches and the application of triage rules to determine which projects to analyze, are found to be useful but not robust.