Decisions taken at the early stages of a project have a far reaching effect and set the tone for the remainder of the project. In the early stage, creative solutions can slash delivery times in half and cut costs dramatically. However, once development is under way, it is rarely possible to effect savings of anything but a few percent. Good upfront work also reduces the likelihood of change later, as most changes on projects are actually reactive to misunderstandings over scope and approach...As I read this, a post at Construction Survivor was my immediate reference: Value Management -- the great con trick.
Gerry is right if VM is done when the project is largely committed and all the valuable change opportunities have gone. VM is only of any use done early; and best coupled with project planning and definition workshops with the main project users participating. The NSW Value Management guide, as an example which I've made use of in many project development workshops, makes the point on this quite succinctly.