When you run a usability test, you should already have some information on issues with your site or design. This can come from focus groups, surveys, log analysis, customer feedback, and other user research you performed earlier in your design cycle. These should give you a good idea of where to focus your resources. Its a mistake to ignore this data and start from scratch.
Example: From looking at site logs, you know that an abnormal number of customers are leaving from the "View Shopping Cart" page. The logs don't necessarily tell you why this is happening, so you should look to see if the abandonment is caused by a usability issue.
