The main focus of software testing is to prevent issues from reaching end-users. To achieve that, smart test managers and testing teams employ several techniques to increase test coverage and avoid bugs slipping into production. Below are some simple yet effective practices to take into consideration for a solid test strategy.
Testing should be planned ahead, right along with software requirement analysis. As most Agile frameworks and methodologies advocate, the testing team should be involved in the requirement gathering phase as early as possible. This helps build context for the user experience as well as the core purpose of software or feature.
While development teams work through requirements, testing teams should work to create testing plans and test cases that align with acceptance criteria. Dividing specifications into smoke and detailed feature test cases helps create layers to validate quality as the software moves through the testing cycles. Very often, the questions asked by the testers during the requirements analysis will lead to further improvements before any code is developed.
Well-structured test cases allow testing teams to approach testing in a systematic manner, therefore eliminating chances that any acceptance criteria are missed in the testing process.
One of the techniques widely used to structure test cases (and also to prioritize bugs) is the probability/impact matrix. When creating test cases, it’s important to have a probability and impact chart which analyzes the new software requirements and acceptance criteria for regression in the larger software system. This helps foresee potential bug-heavy areas post-development. The probability/impact matrix is a key component to Risk-Based Testing (RBT).
Take for example a new feature that allows users to upload a profile picture on a web application at the time of signup (optional, not mandatory) – the probability/impact matrix for this feature will look at:
- impact on signup if the user doesn’t upload a profile picture
- impact on signup in case the server returns an error on profile picture upload
- impact on signup if the user uploads a non-standard size picture, etc.
If any of these test scenarios result in a failed signup, the impact will be high. This impact then needs to be assessed for the probability of occurrence i.e. how likely is this scenario to occur. The more probable and high-impact a scenario is, the more test cases (negative and positive) should be developed around it. These areas are also prime candidates for exploratory testing in later test cycles.
For insights on how to best use risk-based approaches to optimize testing, read our QA expert John Ruberto’s explanations and advice.
Sharing Test Cases With Development Teams
More and more, agile teams have begun to share test cases with developers at the start of development or as soon as the test cases are ready. While it is never required for developers to execute all the test cases provided by the testing team, they do provide insight and detailed information on how the software is expected to behave. Reviewing the test cases with developers help both teams – developers can help testers identify code level impacts which may not be apparent from functional analysis, while testers can help developers identify scenarios and data points that need to be provisioned for in the software design.
Structuring and planning tests help maximize quality from the very start of the software development process. With every testing cycle, the software becomes stable in the area covered by the test cases. However, it is never wise to assume that further testing is not required. A round of exploratory testing – which draws on data from earlier test cycles and a user interaction heatmap (or equivalent data) to test corner cases or complicated scenarios – should be performed.
For the profile picture upload example mentioned earlier, this can include testing what would happen if, for example, the image uploaded successfully but the user put the desktop to sleep and returned the next day – would sign up continue successfully?
Asking testers to switch test areas with each other (so essentially cross-test) at the end of an exploratory test cycle will help identify more hidden bugs.
All in all, the best way to improve quality during testing is to never become complacent with any particular method. Testing teams should continue to learn from each release and work on improving test coverage for the next one.