Timing or Other Benchmarks

You may want to use time as a criterion for success or as a benchmark. If you do set benchmarks that are based on timing, it's recommended you do this very thoughtfully and under just the right circumstances. For example time-on-task is a good measure for validation/summative tests, but it is rarely appropriate for early exploratory or formative tests. It is inadvisable to measure time-on-task if you're also asking participants to think aloud, because doing so typically slows task performance. For more about benchmark timings, see

the Benchmark Timings sidebar. If you don't want to use time as a benchmark you could use error rates; for example, completing a task with no errors of any kiding