Attention on social media depends more on how you express yourself than on who you are.

Attention has become a vital form of capital in the digital age, yet the mechanisms underlying its allocation on social media remain poorly understood. Using a nationally representative, online and offline-integrated dataset of a Generation Z cohort in China, we provide large-scale evidence on the determinants of success in attracting attention. Our findings reveal that 'how you express yourself' (using various emojis and expressing multiple emotions) is more influential than 'who you are' (in terms of gender, education, family background and personality traits) in attracting attention on social media. Further analysis confirms a causal effect of the variety of emojis and types of emotions on attracted attention, while simulation processes using agent-based models suggest that empathy evocation is the primary underlying mechanism. We also show that the mode of expression is largely independent of individual characteristics and that the attention gained from highly appealing expressions is easier to acquire than to sustain, as it is highly sensitive to changes in expression modes over time. Overall, our research identifies three key features of attention capital allocation on social media: low alignment with traditional resources, considerable manipulability and ease of acquisition but difficulty sustaining it over time.