Polarization is a well-documented phenomenon across a wide range of social issues. However, prevailing theories often compartmentalize the examination of herding behavior and opinion convergence within different contexts. In this study, we delve into the micro-foundations of how individuals strategically select reference groups, offering insight into a dynamic process where both individual opinions and the network evolve simultaneously. We base our model on two parameters: people's direct benefit from connections and their adaptability in adjusting their opinions. Our research highlights which conditions impede the network from achieving complete connectivity, resulting in enduring polarization. Notably, our model also reveals that polarization can transiently emerge during the transition towards consensus. We explore the connection between these scenarios and a critical network metric: the initial diameter, under specific conditions related to the initial distribution of opinions.