A mutual fund's idiosyncratic activeness (IDA) is the contribution of idiosyncratic shocks, relative to systematic shocks, in explaining the variation of its active returns. This novel measure of the source of activeness of mutual funds is informative about the way they strive to beat their benchmarks, and delivers new cross-sectional predictions of fund performance. We show that funds whose activeness is mostly idiosyncratic (high-IDA) significantly outperform those whose activeness is mostly systematic (low-IDA), and uncover that being more active only leads to better performance if IDA is sufficiently high. By further decomposing a fund's systematic activeness, we are able to identify funds that rely on smart-beta strategies and show that those that focus on multiple factors (within a quarter or over time) tend to underperform those that focus on a single factor.