In the past in my current position we reported results in terms of failure rates. Our release critera was based on failure rates. Our ISO internal auditors loved this. However, these numbers were more than often misleading. They were not a true measure of the quality of the product under test. Some features with a small number of test but with 1 failure showed a failure rate of 10% where larger features with more failures showed a much lower acceptable failure rate. However the larger feature would be the one that we were concerned about but the numbers showed it was acceptable.