Comprehensive post Steve. You've outlined also some contrasting approaches to measurement.
At one level is having 'discrete proxy metrics' that are relevant to a niche and as you rightly mentioned, unlikely to be transferable outside that niche. We could call these as bottom-up driven metrics whose tracking is critical for any organization.
At another level are the metrics to measure a larger sector or an industry like the kind you described with IRIS/ Pulse which will take time to develop. Unless an entire sector (or part of a sector) adopts the same metrics this will just be an overhead for the reporting organization. We could call these top-down metrics and if there are a lot of these, the emergence of a standard will become challenging. There are a few of these emerging, be it through privately driven organizations (like B-corps) or government driven (like the reporting in the new Form 990s or with the new legal structure for social corporations ... i forget what it is called)or consortium driven (like the Global Reporting Initiative).
Is there an intermediate step that leverages what organizations are already producing? CharityNavigator seems to have done something similar where they take data reported by the organizations in their Form 990s and then based on their areas of operations start building comparable metrics. In such a case you're not comparing a cancer research organization with a food distribution program. Once organizations see value in the data and comparisons being made they will be more proactive in making the linkage between the 'bottom-up discrete-proxy metrics' and the 'top-down sector metrics'. What are your thoughts on such a middle-ground or medium-term opportunity for some sort of a 'meta-data service provider' for social impact?