Remove personnel incentives to pass program milestones in ANY acquisition program. Personal grades and promotions should be on successful management of a program, not if it passes milestones. Successful management can include stopping a program and rolling it back to a previous milestone if so warranted.
If a program is created with bad requirements, has poor implementation, has insufficiently developed technology, the current personnel running the acquisition program and in charge of acquisition programs are personally incentivized to cause the program to succeed in order to garner a successful job evaluation. The culture cannot say no to a program for fear of a bad evaluation, the program or platform losing money, or loss of a job. The process to STOP a program at a milestone is almost never used because every person involved is incentivized to pass a program until congress is forced to cancel it.
If leaders are directed to evaluate their acquisition subordinates on running an acquisition program without incentivizing success or failure, more programs may be rolled back to previous milestones and more complete products may be developed. More money may be saved by developing a better, more complete product, or requirements may be refined before the program has progressed to production. Incentivize acquisition personnel to fairly judge a program instead of passing this program for fear of job, evaluation, or program money and better, more cost effective products would develop.
Programs halted and recycled through milestones in this manner may cost more in the development phase, but would cost significantly less in the sustainment phase when incomplete supply chains, poor reliability, poor implementation, and other factors cause the new product's costs to far, far outstrip any benefit saved during the product's development.
The James Webb Telescope could be used as an example. Did it need better requirements? Technology? Which milestone should it fail?