What to Test is Clear – How to Decide What NOT to Test?

You are here

what to testWhen there is a lot to test, but cost or time is insufficient, the number one problem for the testing team is to decide not what to test but what to leave without testing. If the project is really complex and involves many features, there are a lot of possible testing strategies and multiple test cases to execute, but if you physically cannot manage to do all of these, you should first carefully think over what is the most and least important in your product. Most likely, the quality of the product and therefore its success and the company’s profit will depend of what you choose to test. If the matter was about what to include in testing, there would be little doubts. For instance, you cannot afford to skip the UI in favor of testing backing database. In case when you’ve got to decide what to exclude from testing, everything’s much more complex and there is no single scheme for any project. So, to give you something for orientation, consider the following exclusion criteria that may help you to decide rationally.

  • You need the broad testing coverage. It is usually better to embrace every aspect a little than skip one part and focus totally on another one.
  • It is wiser to include in the test scenario those test cases that the users are most likely to interact with in the first place. That is, you should search for intersections of the features with user activity. It won’t matter if there is a bug in the place where users will never get, but it may cost very much if a lot of users come across the bug in the most interactive part of the product.
  • Analyze your product for the most problematic areas in case something goes wrong. The risk is higher when the data is concerned, so this aspect is the most important than the drawing, for example. Again, the legal requirements are another significant area, so make sure to test them in the first place.
  • It will be also rational to test cheaper parts in terms of servicing if a bug appears. One of such important areas is client-server architecture.
  • This criterion is not about the usefulness or importance, but more about cost-effectiveness. Logically, if some test case is expensive to execute, it is better to choose several other test cases that are cheaper.
  • Test cases which cover new features bring more incremental use than those covering partly old and partly new features, though it is recommended to incorporate both.

The main idea of effective testing is to make an informed decision before getting started. It will not bring any good if you start carrying out test cases randomly. Estimate your chances with the presented criteria and test software more efficiently!