The rapid acceleration in the use of digital tools and platforms in education, driven by the COVID-19 pandemic, has also impacted Massive Open Online Courses (MOOCs), which are now available globally in various formats. An increasing number of individuals engage with these courses to enhance their knowledge, skills, and competencies for personal and professional development, benefiting from their free and easily accessible nature. This trend is also evident in our country, despite the absence of a definitive stance from European institutions regarding the accreditation and legal recognition of this type of education. Contemporary MOOCs increasingly emphasize brevity, conciseness, and reusability in their content design, prioritizing audiovisual materials over text-based resources. This shift aligns with an increasingly fast-paced, flexible, and personalized approach to digital learning. Consequently, training providers face the challenge of ensuring the sustainability of these courses, as they must continuously produce and update materials that rapidly evolve. In this context of fast, flexible, and at times consumer-driven self-directed learning, it becomes particularly relevant to reintroduce the value of socially oriented learning within MOOCs. This necessitates a critical examination of sustainable engagement and participation strategies. This study, conducted by CREMIT (Research Center on Media Education, Innovation, and Technology) and ILAB (Center for Innovation and Development of University Teaching and Technology) at Università Cattolica del Sacro Cuore, aims to: 1. Compare the instructional models of pre-COVID (tutored) and post-COVID (self-paced) MOOCs, highlighting changes and challenges related to the social dimension of learning; 2. Identify signs of community formation within self-paced MOOCs; 3. Identify key conditions for fostering and sustaining a learning community that extends beyond the conventional MOOC lifecycle. To achieve these objectives, the study collected and analyzed monitoring data from 33 MOOCs delivered between 2014 and January 2025. The dataset includes both quantitative data (initial and final questionnaires, learning analytics, and retention metrics) and qualitative data (analysis of forum messaging). The contribution concludes with two forward-looking proposals: a hypothesis for the redesign of MOOCs, incorporating support tools for learner guidance and self-regulation; and the presentation of a new collaborative space dedicated to MOOC-related themes, envisioned as a platform for dialogue, reflection, and the co-design of new educational materials.