Depends on your org and it's goals. For orgs that are engineering oriented with strong business plan, strong requirements gathering, case studying, and minimal

Depends on your org and it's goals. For orgs that are engineering oriented with strong business plan, strong requirements gathering, case studying, and minimal feature selection, it's a boon. For orgs that aren't engineering oriented and use development as a tool to find holes in the business plan, testing anything is usually a waste of time. You're trying to see what the market responds to, rather than actually build something of quality. It comes down to each org's goal and usually the best answer will be somewhere between the middle of the two extremes.