BACKGROUND: Public health policies sometimes have unexpected effects. Understanding how policies and interventions lead to outcomes is essential if policymakers and researchers are to intervene effectively and reduce harmful and other unintended consequences (UCs) of their actions. Yet, evaluating complex mechanisms and outcomes is challenging, even before considering how to predict assess and understand outcomes and UCs when interventions are scaled up. We aimed to explore with UK policymakers why some policies have UCs, and how researchers and policymakers should respond. METHODS: We convened a one-day workshop with 14 people involved in developing, implementing or evaluating social and public health policies, and/or evaluating possible unintended effects. This included senior evaluators, policymakers from government and associated agencies, and researchers, covering policy domains from public health, social policy, poverty, and international development. RESULTS: Policymakers suggested UCs happen for a range of reasons: poor policy design, unclear articulation of policy mechanisms or goals, or unclear or inappropriate evidence use, including evaluation techniques. While not always avoidable, it was felt that UCs could be partially mitigated by better use of theory and evidence, better involvement of stakeholders in concurrent design and evaluation of policies, and appropriate evaluation systems. CONCLUSIONS: UCs can be used to explore the mechanisms underpinning social change caused by public health policies. Articulating these mechanisms is essential for truly evidence-informed decision-making, to enable informed debate about policy options, and to develop evaluation techniques. Future work includes trying to develop a holistic stakeholder-led evaluation process.