Throughout modern democracies, political parties’ decisions on programs, policies, and personnel are the result of intraparty actors’ preferences and discussions. However, these internal dynamics behind such decisions remain largely unknown because of strong data limitations, known as the “black box of intra-party politics”. We argue that motions, i.e., proposed courses of action that are brought before a party assembly for consideration, debate, and/or decision by an individual or a group that are members/affiliates of the party, offer a unique window to reveal the processes and dynamics behind parties’ official decisions. In this paper, we demonstrate the potential of motions with the help of a novel dataset, the Party Congress Motion Dataset (PCMD). It contains motions proposed to annual party congresses (N = 149) from all major parties in the United Kingdom, Germany, and the Netherlands (N = 20) over the period 2017-2022, a total of more than 17,413 individual motions. We use the data to create measures of intraparty preferences, engagement, divisions, and actors’ relative power that are more precise, valid, and dynamic than existing ones. The results show how varied intraparty preferences and dynamics are across European parties and time, and that they are not always following the expected patterns. Methodologically, we contribute to the literature by showing how to extract meaningful and comparable information from large, diverse, and unstructured number of archival documents via a mix of manual and computational methods.