Our online social lives span various distinct circles: family, friends, work peers, and communities. When these circles merge in one online space, tensions can arise; because norms and expectations differ across groups. This blending of separate social contexts into a single digital presence is known as context collapse. But does context collapse undermine users’ experience so much that it can lead to masses leaving the platform? Focusing on ego networks of 10,000 randomly selected users of iWiW, a now-defunct Hungarian social network, the team created a quantitative indicator context collapse based on network structure through two key indicators: (1) Network fragmentation, measured via modularity—i.e., whether communities are distinct vs. overlapping. (2) Social differences between communities, such as age and urban vs. rural composition. Their results shown that Users whose networks were fragmented and socially diverse (e.g., several separate communities with different age profiles or urban-rural mix) tended to exit the platform earlier. This suggests context collapse creates tension when diverse groups with differing norms co-exist but don’t interact. Yet, the highest risk of leaving fell on users with non-fragmented networks where community members were socially similar. This might seem contradictory, but the interpretation is key: network fragmentation itself tends to reduce the probability of leaving, because fragmented users have multiple engagement channels to fall back on.