When we hear about new and innovative methods for managing projects, let's ask some questions:
- What domain and context are these approaches known to work in?
- Have they been applied in any domain and context we are interested in?
- Is that any tangible evidence that the approach actually works outside of anecdotal and personal experience?
The check list above is applicable in large complex development programs, where requirements are emerging, but the needed capabilities have some level of stability. This means the customer knows to some level of confidence what done looks like in meaningful units of measure.
The critical few in this picture are:
- Inattention to budget - if you don't have some notion of what this project is going to cost when you're done, it'll cost what it costs.
- Lack of integrated management system - cost, schedule, resource, technical, and customer management all have to be connected.
- Lack of predictive variance analysis - without forecasts of what is coming done the road, you'll only find out when it is too late. Probabilistic forecasting is a critical success factor once any project moves beyond the small stage. For projects described by Edwin Land, that are manifestly important and nearly impossible, predicitive variance analysis is mandatory.
The others are specific to a project domain and context. For now the critical few are the show stoppers for any complex project, no matter the domain or context.
