The United Kingdom Research and Innovation (UKRI) Global Challenges Research Fund (GCRF) aimed to address global challenges to achieve the United Nations (UN) Sustainable Development Goals through 12 interdisciplinary research hubs. This research documents key lessons learned around working with Theory of Change (ToC) to guide Monitoring, Evaluation and Learning (MEL) within these complex research for development hubs. Interviews and document reviews were conducted in ten of the research hubs. The results revealed that only one hub invested in an explicit visual system mapping approach, and that funder timelines, budget constraints and issues with capacity and expertise limited the application of these approaches across all hubs. In contrast, many hubs attempted to deal with visual complexity by means of ether constructing multiple, nested ToCs, or a conscious simplification of complexity through reducing their ToC towards a straightforward and uncomplicated chain model or spherical model. While the former approach had some value, most hubs struggled to find capacity to support the full articulation of nested ToCs. In contrast, the latter approach resulted in ToCs which lacked detail or mechanism articulation, but which nevertheless were often 'fit for purpose' in ensuring effective communication and coherence across diverse stakeholders and sub-projects. We conclude that in instances where the reporting, funding and management cycles of complex research for development programmes cannot be adapted to properly support learning-based approaches to ToC development, imposing simplicity in the ToC might be fit for purpose. This might also be preferable to more complex visual approaches that are only partially realised.