Severity is a judgment, made by the opener, of the seriousness of the problem as related to the product function.
It represents the impact this problem would have if the customer had seen it. Severity should not change during the life of a problem. Severity will be assigned according to the following criteria:
Severity 1: An unavoidable error which makes the software under test inoperative, including any error which requires a restart of the operating system. A severity 1 defect might also involve a data integrity failure and/or jeopardize test conduct. A failure which gates the testing of an entire line item may also be considered severity one.
Severity 2: A significant functional defect which may invalidate continued testing in the functional area. The defect may mask the further detection of problems in the functional area affected. All software crashes encountered can be considered severity two.
Severity 3: Any minor functional problem or technical inaccuracy in the documentation.
Severity 4: A very low impact defect. Misspellings, color problems, incorrectly formatted screens, bad grammar, and other non-functional problems are examples.