Development asks that the inequity and unsustainability of the widening gap between rich and poor be narrowed, ultimately impacting on households in the most economically excluded communities. Local community-based organisations (CBOs) provide much of the organisational fabric through which development is delivered. Largely resourced by the poorest themselves, many of these CBOs aspire to attracting funds from the development aid industry. In attempting to comply with the rules of these funding sources and compete in funding relationships, organisations become players in the funding game fraught with power imbalance and seemingly contradictory incentives. Neither the funding agencies, intent on disbursement, nor the CBOs in their desire to build organisations and contribute to their communities, seem aware of the true costs of these relationships. Aid funding is complex, operating at numerous levels, across a multiplicity of varied organisations, stakeholders and contexts. Over the last 60 years, the aid industry has evolved complicated and highly engineered mechanisms to manage relationships with funding recipients, including detailed conventions for evaluation. As part of contractual obligation, criteria for success are pre-defined; outcomes are predicted; and targets are projected. Development, however, is not linear or predictable. It is contradictory and complex. Despite objections and alternatives since the late 1980s, ‘conventional’ linear, simplistic rationale has dogged the development industry. The HIV support sector as a focus for funding, capacity building and service contracts from government and international aid agencies, offers rich examples of aid industry dynamics. This research, set amongst small but established CBOs working in HIV/AIDS support in Soweto and Lawley (Gauteng) and Mabeskraal (North West Province), explores alternative evaluation approaches, methodologies and principles, based on grounded evaluation. Two models are tested and compared. Firstly, inward-looking, organisation-based, reflective self-evaluation using Stories and Metaphor. Than secondly, outward-looking, community research using a Most Significant Change approach. The evaluation processes developed help participating CBOs describe success and outcomes against their own criteria. The approaches use narrative, visual and metaphorical formats. The central purpose of the research is meta-evaluation aimed at an effective process using iterative, cumulative action research based on the principles of grounded theory. Meta-evaluation data included descriptions of the processes and the nature of evaluation results. They are analysed using reflection, learning and re-design in an action research cycle. The results provide both practical insights into conducting evaluation, and the principles of effective development in a CBO setting. They demonstrate that grounded evaluation can be used to understand organisational dynamics and programme outcomes. Participatory methods, particularly visual and verbal communication, are shown to be far superior to written communication in this setting. The results demonstrate the mutual compatibility and ethical inseparability of organisation development with evaluation, providing insight into the practice of utilisation-based evaluation. The value of appreciative inquiry and the risks of accusatory inquiry are described. A thread that runs through the results highlights the impact of power, ownership and process use in effective evaluation. The research has also elaborated some of the intractable contradictions and conundrums in development aid. Money carries the power vested in global economics and market forces. In making funding judgements, evaluators purvey the power of wealth inequity: the very power imbalance which itself purports to address. As a development practitioner, an evaluator’s role should be to facilitate pathways out of dependent mindsets. As gatekeepers to financial support, however, their work entrenches distortions in perceptions of wealth and power. These complex interactions of power and ownership demand moderation and compromise. The industry requires investment of greater energy into theoretical, methodological and practical research. Suggestions for such research are included. Without fresh creativity, development and evaluation will remain frustrated forces within an entrenched, self-perpetuating system of inequity and disparity.