In Dhaka, Bangladesh over five million people live in low-income, informal settlements (bustees) with limited access to basic services, secure land tenure and political voice. Whilst collective action among the urban poor is central to accessing affordable services and - when taken to scale - a broader politics of 'redistribution, recognition and representation' (Fraser 1997; 2005), little is known about how Dhaka's slum dwellers organise, and the extent to which this is (or can be) transformative. To deepen our understanding, this thesis utilises collective action theory to examine intra-group dynamics, the instrumental value of groups and broader context of urban governance that enables and/or constrains certain forms of collective action in Dhaka's bustees. Case studies of Community Based Organisations (CBOs) in three bustees are used as a lens to explore how slum dwellers organise to obtain basic services, such as water and sanitation. CBOs are disaggregated into two main types (externally or NGO-initiated and internally or leader-initiated) and sub-types (formal and informal), with three sub-themes; participation (leadership and membership), function (activities and responsibilities) and outcomes (equity and sustainability). A mixed qualitative toolkit, including in-depth observations of CBOs, interviews with CBO leaders, members, non-members and key-informant interviews with NGO, government officials and citywide urban poor groups, reveals the complex relationship between collective action, service provision and urban governance in Dhaka. Two key findings emerge. Firstly, similar patterns in participation and outcomes are observed regardless of CBO type, whereby politically-affiliated local leaders and house owners create, enter and/or use CBOs to address their strategic agendas, and reinforce their authority. This demonstrates that, as opposed to bounded groups, CBOs are in fact nodes of interconnected individuals, some of whom are better able to participate in (and benefit from) collective action, than others. Secondly, although collective action plays an increasingly important role in service provision in Dhaka (especially legal water supply), it is largely practical in nature (i.e. addressing immediate needs). In cases where it is more strategic (i.e. to access land and housing), or both practical and strategic (i.e. obtaining legal water supply to secure land), certain male local leaders seek to benefit over others. In all cases, transformative collective action is constrained. This, it is argued, relates to the broader context of urban governance that enables certain forms of collective action, while constraining others, in Dhaka's bustees. Three (interrelated) spheres of urban governance are identified as particularly important: 1) patron-centric state; 2) risk-averse and market-oriented development sector; and 3) clientelistic society. Whilst existing collective action theory has value for understanding intra-group dynamics, fieldwork suggests that the urban governance context is the overarching factor affecting collective action in Dhaka's bustees. The thesis concludes with potential ways forward.