The efficient delivery of government services to the poor, or Bottom of the Pyramid (BOP), faces many challenges. While a core problem is the lack of scalability, that could be solved by the rapid proliferation of platforms and associated ecosystems. Existing research involving platforms focus on modularity, openness, ecosystem leadership and governance, as well as on their impact on innovation, scale and agility. However, existing studies fail to explore the role of platform in scalable e-government services delivery on an empirical level. Based on an in-depth case study of the world's largest biometric identity platform, used by millions of the poor in India, we develop a set of propositions connecting the attributes of a digital platform ecosystem to different indicators for the scalability of government service delivery. We found that modular architecture, combined with limited functionality in core modules, and open standards combined with controlled access and ecosystem governance enabled by keystone behaviour, have a positive impact on scalability. The research provides insights to policy-makers and government officials alike, particularly those in nations struggling to provide basic services to poor and marginalised. This is also one of the few studies to apply platform theory empirically to the e-government domain.