As climate change impacts unfold across the globe, growing attention is paid toward producing climate services that support adaptation decision-making. Academia, funding agencies, and decision-makers generally agree that stakeholder engagement in co-producing knowledge is key to ensure effective decision support. However, co-production processes remain challenging to evaluate, given their many intangible effects, long time horizons, and inherent complexity. Moreover, how such evaluation should look like is understudied. In this paper, we therefore propose four methodological guidelines designed to evaluate co-produced climate services: (i) engaging in adaptive learning by applying developmental evaluation practices, (ii) building and refining a theory of change, (iii) involving stakeholders using participatory evaluation methods, and (iv) combining different data collection methods that incorporate visual products. These methodological guidelines offset previously identified evaluation challenges and shortcomings, and can be used to help stakeholders rethink research impact evaluation through their complementary properties to identify complex change pathways, external factors, intangible effects, and unexpected outcomes.