In recent years, researchers have shifted their focus to studying the effects of peacekeeping in a geographically and temporally disaggregated manner. One of the factors that is yet to be fully examined is the variation among peacekeeping troops at the local level, and its impact on peacekeeping effectiveness. Specifically, peacekeeping troops greatly vary across two dimensions:\ unit types, e.g. infantry, engineering, aviation etc., and their country of origin. We argue that different unit types greatly increase peacekeepers' specialization in skills and equipment, improving peacekeeping effectiveness in reducing battle-related deaths. However, we also argue that this effect is strongly moderated by cultural diversity among troop contributing countries (TCCs), i.e. linguistic and religious distance, which exacerbates coordination problems among troops. We conduct semi-structured interviews with current and former UN peacekeepers to probe several theoretical mechanisms. We then test the propositions derived from our theoretical discussion using new subnational peacekeeping data, Geo-PKO (Cil et al. 2019). Using matching and a difference-in-difference design, our analyses on UN peacekeeping bases across Africa from 1994 to 2014 shows that force structure at the local level significantly impacts peacekeeping effectiveness.