Every year, organizations around the world spend billions of dollars trying to improve people's lives and protect the environment. Some of those programs work. Some don't. Almost nobody verifies which is which.
The structural reason is the same across domains. Improvement processes are designed as loops — measure, act, measure again. In practice, every project gets optimized for the measurement phase. The actual work of acting on what was measured, and learning what worked, is treated as an afterthought. The feedback loop collapses, and has for seventy years across every framework that tried to build it.
Government is where the gap is most visible. The most ambitious agentic government systems in the world — Ukraine's Diia, Singapore's LifeSG, Estonia's AI Stack — deliver services to citizens at unprecedented scale. None of them ask whether the services produced outcomes. The infrastructure for delivery exists. The infrastructure for accountability does not.
SDOH builds the half that's missing.