When to deliver
Sooner or later, the final output of the project, including all its deliverables will be delivered to the customer and the end users. The when and how of this depends on many factors:
- Predictive projects: Because of the more-or-less linear development in predictive projects, their outputs are usually not usable until the end of the project. (Think of a project for building a bridge.) Therefore, the output is usually delivered in one go, at the end of the project. On the other hand, some predictive projects may have major phases that end at different times and create usable subsets of the final output, and therefore, it may be possible to deliver them in a few steps instead of all together at the end.
- Adaptive projects: Adaptive projects create usable increments of the final output throughout the project to collect feedback and guide the way forward. These increments can be released to production and used by real end users, but when it’s not possible (because of the operations difficulties), they will be given to subsets of end users, or even representatives of them. In this way, their delivery can be continuous.
An increment of the output is a set of deliverables that are usable by the end users; i.e., every increment is a set of deliverables, but not every set of deliverables is an increment. Projects can’t be adaptive unless they have incremental delivery, because useful feedback can only be generated by something usable, and not with other sets of deliverables that may be too abstract for typical users to understand.
Next: Delivery risk