Group collaboration is becoming increasingly important these days. In order to make group collaboration as effective as possible, it is crucial to understand the determinants of successful group collaboration. Researchers have focused on studying group collaboration in laboratory settings. Lack of research remains for real-life settings. In this work, genuine data were collected from unstructured out-of-class meetings of the Sloan Fellow program, an immersive MBA program for mid-career mangers, and the data were analyzed by means of advanced regression analysis. A wide range of different types of data (participation, demographic, personality, performance, and satisfaction data) was combined in order to get a broad understanding of group collaboration. The data revealed that Asians tend to speak less frequently, extroverts tend to speak more frequently, and women tend to have shorter turns. They further revealed that groups with a higher level of conscientiousness or openness to experience have a more balanced participation. Individual participation was shown to have a positive correlation with performance as well as satisfaction. Against expectations, no correlation was observed between the balance of participation and performance as well as satisfaction. Real-time feedback was shown to support group members adapting their behavior on average. More data will be required in order to prove this finding though. Against prior findings from laboratory settings, a novel hypothesis postulating that even participation is not necessarily beneficial in real-life settings was created. Although the first findings support the hypothesis, further investigation of the hypothesis will be required. The present work uncovers insightful patterns of social interactions and provides a deep understanding of group collaboration. The findings are particularly valuable because they can be used in order to improve the effectiveness of group collaboration.