You're analyzing growth metrics for a ride-sharing product and notice a few days with unusually high completed rides and new rider activations. You need to decide whether these points reflect real product behavior, logging issues, or one-off operational events.
How would you identify and handle outliers in a growth analysis?