In an approach to determining speech effectiveness, one or more computer processors receive speech input. The one or more computer processors determine, based, at least in part, on the received speech input, a first speaking mode. The one or more computer processors detect, based, at least in part, on the first speaking mode, at least one aspect of the received speech input. The one or more computer processors determine, based, at least in part, on the first speaking mode and the detected at least one aspect of the received speech input, whether a rate of occurrence of the detected at least one aspect of the speech input exceeds a pre-defined threshold. Responsive to determining a rate of occurrence of the detected at least one aspect of the speech input exceeds a pre-defined threshold, the one or more computer processors notify a user.