There may be a better word for this, but I picked this from terminology I am familiar with.
As you get to the point of diminishing returns on an effort, you will find that improvement of one performance parameter results in a detriment in others. So the overall product is different, but not noticeably better. This means that the cost reduction effort is no longer useful and should be terminated.
For instance, you changed the electrical approach and reduced the cost of the electronics 5%. But now the Printed Wiring Board is larger, and will cost 4% more. The overall result is very different with a higher risk, and not much cheaper.
When you begin to detect dithering, you know you are at the point of diminishing returns.