Abstract: Variety of approaches to optimize performance based on behavioral and situational signals are described. A productivity service initiates operations to optimize performance upon receiving a behavioral signal and/or a situational signal associated with a user. The behavioral signal and/or the situational signal are analyzed to infer a usage pattern associated with a component of an application. Next, a failure of the usage pattern is detected in a comparison to an utility threshold. In response, the application is instructed to interrupt an execution of the component.

Abstract: A computer-implemented technique is described herein for customizing user experience on a per-user basis. In some implementations, the technique involves: collecting feedback information from a user pertaining to the user's interaction with a program feature; determining a modification to be made to the program feature based on the feedback information; making the modification; notifying the user of the modification; and receiving the user's post-modification assessment of the modification. The technique enables a program provider to quickly modify program features in an automated manner, with no (or reduced) effort by human developers, and at consequent reduced cost. In addition, the technique improves the users' experience with the program features since the program modifications are tailored to each user's preferences.

Abstract: Variety of approaches to optimize performance based on behavioral and situational signals are described. A productivity service initiates operations to optimize performance upon receiving a behavioral signal and/or a situational signal associated with a user. The behavioral signal and/or the situational signal are analyzed to infer a usage pattern associated with a component of an application. Next, a failure of the usage pattern is detected in a comparison to an utility threshold. In response, the application is instructed to interrupt an execution of the component.

Abstract: A computer-implemented technique is described herein for customizing user experience on a per-user basis. In some implementations, the technique involves: collecting feedback information from a user pertaining to the user's interaction with a program feature; determining a modification to be made to the program feature based on the feedback information; making the modification; notifying the user of the modification; and receiving the user's post-modification assessment of the modification. The technique enables a program provider to quickly modify program features in an automated manner, with no (or reduced) effort by human developers, and at consequent reduced cost. In addition, the technique improves the users' experience with the program features since the program modifications are tailored to each user's preferences.