International organisations have measured adult learning participation since the 1990s, using surveys like the Adult Education Survey (AES) and the Programme for the International Assessment of Adult Competencies (PIAAC). These surveys employ a 'triadic' classification of learning activities -- formal (FAE), non-formal (NFE) and informal learning (IFL) -- to compare participation rates. However, historical and conceptual advancements have highlighted that these categories often overlap, challenging their validity and reliability. Despite this, the triadic classification remains in use, raising concerns about the accuracy of participation data. This article critically examines the continued use of the triadic classification, arguing for a conceptual shift to better reflect the contemporary learning landscape. We review the evolution of the triadic conceptualisation, particularly within the EC's Classification of Learning Activities (CLA), and critique its practical application in the recent AES (2022) survey. Based on our critique, we propose two alternative scenarios: (1) treating learning categories as fuzzy, overlapping concepts, and (2) conceptualising learning activities along a continuum of formality/informality. We conclude by discussing the implications of these scenarios for future data collection and measurement practices, advocating for instruments that adapt to the evolving digital and blended learning landscapes accelerated by the pandemic.