The economics book Design Rules: The Power of Modularity suggests orgs use something like NPV to assemble as a basket of competing efforts to tackle problems. The strategy of prematurely picking winners, come hell or high water, always struck me as sub-optimal.