Money For Nothing
... you are developing under a Product Backlog ordered using High Value First and know that customer value is dropping over time.
✥ ✥ ✥
You don't want to spend time doing development that costs more than it is worth. Because of High Value First we know that value to the customer drops monotonically over time. The customer is interested in using our product to increase their own value; they are our next link in the value stream. We could try to do planning up front to determine how much to build, so that we don't over-produce and create waste for the customer, but that's difficult from the beginning. On the other hand, we can identify the decision points up front: they are the Product Backlog Items. We can predict that if you deliver High Value First you will eventually come to a point of diminishing returns.
Therefore: Stop development when the cost of continuing exceeds the benefit that the enterprise will receive from the development.
✥ ✥ ✥
In addition to making sure that customer value isn't decreased, also make sure that stopping development at these points won't cripple your ROI. It is best that you ensure that each deliverable (sprint) has positive ROI so that the customer can't put you at risk of losing money on the life cycle development.
See also Change for Free.
Author: Jim Coplien (from Jeff Sutherland)