repave first and ask questions later post.
There's a certain validity to the idea that if you estimate a feature later, it might be cheaper. After all, you might find that there's a new platform or technology or you've already built it in the interim somewhere else in the company. It's a longer term approach to the idea that you stay agile because things might change of which you'd like to take advantage. But sometimes the business wants you to re-estimate every two to three months because they really want the feature, but they can't come to terms with the cost. And there's magical thinking afoot that the very act of re-estimation will cut the cost to a third or a quarter of what it was originally scoped. What should really be happening is a coordinated, cross-functional, meeting that prioritizes stories and determines what can be done within the cost allowed. And if a minimal unit of functional work can't be achieved, then it should be pushed off considerably, not allowed to chip away development time over and over again.
Snarky: We're hoping the fifth estimate is the charm.
Title: Have you ever seen a groundhog check for his shadow more than once.