Knowledge Content Library
Application of Attrition to Resource Forecasts
Presenter: Bill Reid, GlaxoSmithKline
Presented at the 2003 DAAG Conference in Houston, Texas.
Abstract: Planning resource allocation for a high risk portfolio is not straight forward. A proportion of the predicted workload will never be done, as some projects fail through time. It's normal to take on more work than can be done with the available resource and rely on attrition to keep the workload manageable. Traditionally, managers have used a simple factor to adjust portfolio resource forecasts down. However this may not be successful: when you have a small portfolio of projects, when you have a lot of project decision points all occurring at the same time, when you have no project decision points for a stretch of time, when you are unusually successful, when you are unusually unsuccessful, when you have an unusually large fraction of your resource in one project.
Applying estimated attrition on resource forecasts in a manner which models how it actually happens could provide a better answer. This presentation presents a tool which does just that by applying the uncertainty nodes with estimated confidence of success at the project decision points over a portfolio. It uses a Monte Carlo simulation to predict "attrition adjusted" resource requirements along with confidence limits.
Keywords: modeling modtree, porfolio decision analysis portda portanal, simulation mcsim