The idea is to maximize the usefulness of testing by concentrating on the most important aspects of the system and, therefore, on the areas where defects represent the greatest risk to user acceptance. A further extension of this approach is to also consider the technical complexity of requirements because complexity tends to increase the likelihood of defects. The problem with testing only functional requirements is that nonfunctional requirement areas, such as usability and security, which are important to the overall quality of the system, are ignored. Increasing the efficiency of testing by automating test execution is a good idea. However, by itself, this approach does not ensure the appropriate targeting of test coverage and so is not as effective an alternative. Retesting only defect fixes has a considerable risk that it will not detect instances in which defect fixes may have caused the system to regress, i.e., introduced errors in parts of the system that were previously working correctly. For this reason, it is a best practice to undertake formal regression testing after defect fixes have been implemented.