I want to start with just saying that if you only measure the performance based on time on task etc you will not get the correct picture of how sucessfull the web site is. The most important thing is to look at how the visitor perceive the usefulness. People could have a pretty lousy success rate but still feel that the web site is helping them better than other web site. User interface engineering have studied this and wrote the following article about actual and perceived speed.
If you want more ways to measure the performance, you could use the WebQual framework which measure the web site quality based on Usability, Empathy, Trust, Information and Design. It is developed by University of Bath and the University of East Anglia. You could read about it and copy the methodology here
Constantinides did a meta study to find out what elements did influence the user experience and divided the elements into The Marketing mix, Aesthetics, Trust building, Interactivity and Usability. (The impact of web experience on virtual buing behaviour: an empirical study, Journal of Customer Behaviour, 2005)
All these elements affect the perceived usefulness of the site by the visitor.