This paper investigates how online platform search is used to collect precise information about individual retirement options. We study participants’ log-in data from the
largest pension fund for all government employees and workers in the education sectors in the Netherlands. The fund’s portal updates daily how much pension each participating civil servant in the Netherlands has accrued, and provides a fine-grained view on pension information search behavior of individual workers at all age categories. A 2019 pension agreement, which accounted for a change in the state retirement age, allows for a Tobit RDD analysis of observed online directed search time of the fund’s participants. We find that participants start devoting significantly more search time before the official state retirement age to look into their personal situation only as soon as the possibility to exercise their option to retire early becomes realizable.