By Amrita Jain
Ever wonder how much testing is enough testing? We may be confident through many phases of testing on a project, but there is still no universal criteria to finish up with this ongoing activity. A user can be versatile in the roles of testing by executing functional/non-functional tests to judge scalability or robustness by performing environmental testing for an existing or fresh application. Testers can explore the application, think up new tests behaviorally as an end user, and run them. Even if the user finds no new bugs, this cannot be a set mark for the complete testing, so you don’t want to rule out that there could still be some undiscovered errors.
Complete testing is impossible for several reasons:
- All the inputs to the application can’t be tested.
- All the combinations of inputs to the application can’t be tested.
- All the flows through the application can’t be tested.
- All other potential failures (like those caused by UI design) can’t be tested entirely.
When can you close the QA (testing)?
The criteria for closing the QA depends on the individual product, culture, and environment. Each product is unique and has its own criteria for closing QA. Specialists have a saying: “It’s more about good enough than it is about right or wrong.”
Factors guiding the finish line of QA
If the tester covers the optimized test coverage (number of tested items divided by total number of items to be tested multiplied by 100) and when all the planned tests that were developed have been executed and passed, testing can be finished.
Defect Detection Rate (DDR)
The number of defects found during a test phase divided by the number found by that test phase and any other means afterward is called DDR (defects found by this testing/total defects including those found afterwards). Testers can stop testing when DDR falls below a specified level with enough test coverage. If bug rate falls below a certain level, testers are not getting any blocker/critical/major bugs.
Early testing is relatively simple and cheap. At some point, the cost of finding the “next” defect exceeds the expected loss from that defect. So, testers can finish testing when marginal cost is achieved.
Depending on the market conditions, there may be very logical reasons for shipping the product before finding all bugs, as the first on the market is often the winner. It may satisfy the demands of many users and bring serious benefits to the company. These profit constraints may consequently lead to QA closure.
QA sign-off/closure activities
The completion of QA should always be followed by a formal QA closure process. QA closure activities include the following major tasks:
- Checking which planned deliverables have been delivered.
- Documenting the acceptance of the system.
- Recording the resource management during the project.
- Archiving testware, tools, test environment, and test infrastructure for later reuse.
- Logging the risks and challenges involved in the project.
- Analyzing lessons learned to determine changes needed for future releases and projects.