The "magic number 5"

Common practice holds that 80% of usability findings are discovered after five participants. Recent findings from web testing indicate that a much larger number of participants is required to get results and that independent teams testing the same web-based product do not replicate results. How many users are enough for web testing?