One of the critiques of personas is that the underlying data that they are based on may stale, requiring further rounds of data collection. However, we could find no empirical evidence for this criticism. In this research, we collect monthly demographic data over a two-year period for a large online content publisher and generate fifteen personas each month following an identical algorithmic approach. We then compare the sets of personas month-over-month, year-over-year, and over the whole two-year period. Findings show that there is an average 18.7% change in personas monthly, a 23.3% change yearly, and a 47% change over the entire period. Findings support the critique that personas do change over time and also highlight that changes in the underlying data can occur within a relatively short period. The implication is that organizations using personas should employ ongoing data collection to detect possible persona changes.