What should not be automated?

Before you start test automation, you need to consider the time, effort, and resources required for automation, so that we can get the required Return on Investment.

Below are some factors to consider when deciding which manual tests need to be automated and which manual tests should not. As the old saying goes, just because something can be automated doesn't mean you need to. Do not automate tests that cannot be 100% automated, unless you save a lot of time and effort. Also, to get the most benefit from automation, tests need to provide predictable and reliable results for validating the output.

Here are a few pointers to assist discover right applicants to test cases which should not be automated:

1. New feature tests
2. A test that runs only once. The only exception to this rule is that if you run the test even once on a very large dataset, it makes sense to automate it.
3. Usability tests (a test that requires the user to test how user friendly the application is).
4. Tests that need to be run as soon as possible, for quick feedback for any decision making
5. Exploratory Tests: Tests that require testing based on domain knowledge/expertise
6. Intermittent tests. Tests with no predictable results will generate more noise than the value.?
7. Tests that require visual confirmation
8. Tests that take more time in automating than in executing manually.
9. Tests that have low return on investment
10. Complex tests that require human skill and knowledge while testing

Today, testing is the most important part of software development. There are many benefits of both manual testing and automated testing. So, it’s important to consider both approaches while designing the QA strategy. We need to plan tests so that we are able to give both width and depth coverage to the application. The end result is that a good quality product should be released in the market, while optimizing both time and cost.?

?Happy testing!