Must-solve, long shot things like GPS, Internet, Modern Weaponry require resolve behind it to make it happen.
I am sure those programs had hundreds offshoots trying things out to solve problem, to see if it works, within the umbrella of that major effort e.g. GPS.
ROI measured in terms of money results in drug companies chasing chronic pills than cures. That need not be the only measure.
Then again the Great Retail Apocalypse of 2017 may lead to to think otherwise. US retail is massively overbuild, most of that because the companies are allowed to take on tens or hundreds of millions of debt. This debt acts as a buffer "It's only bad now, it will be better soon, we promise". By the time feedback signals give a clear message that the investment will never be repaid investors are overextended and losses occur.
It really doesn't matter if you are a responsible company that takes on low/no debt if a significant number of your competitors take on debt, and their investors allow it. "Soon" is highly overrated. Amazon is going to make lots of money 'soon' unless they are wrong about future market conditions, yet their investors have given them a pass at reinvesting most of their money for growth. Entire industries can commit irrational actions for long periods of time and have plenty of analyst attempt to rationalize the situation.