Key points are not available for this paper at this time.
We observed the same task executed by 49 users on four production web sites. We tracked the rates of discovery of new usability problems on each site and, using that data, estimated the total number of usability problems on each site and the number of tests we would need to discover every problem. Our findings differ sharply from rules-of-thumb derived from earlier work by Virzi1 and Nielsen2,3 commonly viewed as "industry standards." We found that the four sites we studied would need considerably more than five users to find 85% of the problems
Spool et al. (Sat,) studied this question.