|
|
|
|
|
|
|
ogy and each phase of clinical development, and so forth. Then, the overall probability of success is determined from the product of these individual assessments. |
|
|
|
|
|
|
|
|
The questionnaire must be designed to ensure that risk factors are not double counted, so it is important to emphasize that each expert should consider the probability of success at that phase, independently of all others. |
|
|
|
|
|
|
|
|
A second approach is to ask for a single assessment of the overall probability of success. Although this approach has the advantage of greater simplicity, quality is likely to be diminished in two ways. First, thinking may be dominated by one feature which inappropriately colors judgment of the overall value. Secondly, determining probability of success at each key stage allows reviewing the sensitivities of particular projects and the portfolio to each development risk. Thus, if a number of projects are sensitive, for example, to scale up concerns, this may suggest some need for particular attention to that area. |
|
|
|
|
|
|
|
|
Of paramount importance to successful portfolio review is transparency and buy-in to the source data. In general, project teams generate the source data but these must be reviewed by line management to ensure their support and to take advantage of their knowledge across the portfolio in their area of expertise. It is likely that the project team alone has the necessary in-depth knowledge to understand all the issues on their project, but this must be complemented by line management's ability to look across all the projects and spot inconsistencies or anomalies. This joint approach also ensures full discussion and understanding of assessments to iron out excessive optimism or pessimism. |
|
|
|
|
|
|
|
|
Commercial and resource/cost data are collected by a similar approach, and an initial assessment generated and agreed to by the project team which may be moderated by line management review. |
|
|
|
|
|
|
|
|
This sections considers different ways to analyze input data for useful outputs. |
|
|
|
|
|
|
|
|
There are fundamentally two approaches to reviewing data from the portfolio analysis process. The first is the pure Delphic approach which eschews analysis in favor of debate among experts until a consensus is reached on key issues. The second, this author's preferred approach, favors expert review of data and information derived from a decision analytical approach. The analysis considers probabilities of success, value, and costs. |
|
|
|
|
|