Strategies to Define which Tests Should be Automated

Test Case & Its Importance

Automation testing and its benefits

Automation testing is the use of automation tools to carry out your test cases. This type of testing supports manual testing in doing repetitive testing tasks and other tasks too demanding to perform manually. An effective automation testing pipeline improves testing versatility, enhances software efficiency, and leads to great ROI.

What is the test case?

A test case is a set of documented preconditions that helps a tester know if the software functions correctly. You can think of it as a guide to verify whether a feature under test meets the requirements. A test case may contain one or more test scripts, and a group of test cases is called a test suite.

You must decide the right test case for automation because :

  • It provides good testing coverage for the application.
  • It reduces maintenance and software support costs.
  • It ensures improvement in software quality.
  • It enables the tester to think the process through and carry out the test from the best angle.

How to Choose which Test Case to Automated

There are certain factors to consider before deciding whether to automate a test case or not. These factors include:

  • The complexity of test case – For the tester to know the test case complexity level, they should consider whether tests need to run against multiple configurations (e.g., different OS or Browser combinations) or if tests need to use different data for each test run (e.g., data-driven). Test cases that require these arrangements have to be automated to save the team’s time and effort.
  • The cost of automation vs. the cost of manual testing – After checking the test case’s complexity, the team should analyze the ROI. Test cases that provide a high ROI if automated should be considered for automation.
  • The outcome of automation – What is the intended result of automation? Will it be complicated or easy to define? If the test case does not have a defined outcome for testing, such test cases should not be considered for automation to minimize false results.
  • The downside of automation – The team should ensure that downside of automation won’t overweight the value (or ROI). The downside here includes implementation cost and the maintenance cost while the value consists of the time, effort, and human resources saved.

Test Case to Automate

In order to automate the right test cases, the above-listed criteria must be carefully analyzed. Nevertheless, some test cases should be automated to save time and effort. They are:

  • Regression testing – Tests such as sanity test, smoke test, and regression test should always be considered for automation. They usually require testing in every build and release of the software application. ( Learn more about Regression Testing)
  • API, stress test, load test, performance test, cross-browser test – Automation is the answer to these tests. With automation, testers can access essential parts of the application under tests such as functionality, performance, security, and reliability. Also, automation will help testers achieve the desired results in less time frame.
  • Security Scanning – Test cases with potential security issues in every new feature should be automated to reduce the risk of human error or bugs. Also, tests are known for security weaknesses, and code behaviors, such as encryption ciphers and SQL injection flaws, should go through automation.

But how to develop an optimal automated testing strategy? Check out this practice!

Test Case NOT to Automate

Some of the test cases that should not be considered for automation are:

  • Exploratory test – An exploratory test allows testers to check a system and get ideas about what to test during execution. This test case cannot be automated because, as the name suggests, it is mostly about investigating and learning about the software for the tester. Therefore, it is for testers’ curiosity and doesn’t have a specific testing benchmark.
  • UI test – UI test is done on software that users will come in contact with. It covers screen readers, fonts, menus, colors, and many more. But software testing cannot portray or predict human reactions to this part of the software. Therefore, this type of test needs to be done manually to better testers’ chances of understanding real-life user reactions.
  • Anti-automation test – Anti-automation test cases are for features that require input from human users. Examples of anti-automation features are CAPTCHA, website forms, web-based SMS messaging platform, etc. Automating these test cases will leave it more susceptible to automation attacks (e.g., spams, hacks). Hence, they should be tested manually.
  • GUI verification – This test helps verify the software application’s functionalities to ensure it delivers user requirements. Unlike UI testing, which checks how functions like fonts, menus, colors, etc. feel or look, GUI verification confirms if they work as intended. Manually conducting GUI allows testers to quickly adjust to user requirements – which is usually change-prone.

How to Measure if You Choose the Right Test Cases

Measurements for the right Test Cases

Some criteria to evaluate your test case result to assess if you should continue to automate this test case in the future are:

  • The number of complaints reported by end-users
    The end goal of the testing is to satisfy end-user requirements. Therefore, tracking users’ complaints or reviews (positive and negative) is a key way to measure your test cases’ effectiveness
  • Severity of defects
    Bugs and other issues across the project are another way to evaluate test cases for the future.
  • Number of downtime or system outage
    You can also measure if you choose the right test cases by tracking the frequency at which the product experiences disruptions. That is, how often it shuts down or the duration of downtime events. These data will provide an insight into the software and the bug affecting it.
  • Automation ROI
    This is a significant measure in terms of the test’s increased efficiency, cost-saving, and the obtained outcome. The expected ROI should be measured against the obtained result to prove the test case worthy of automation in the long run.


The fact that Automation is beneficial for testing projects doesn’t mean that testers should automate the whole testing process. Testers only need to choose the right test cases for automation to fully enjoy the process’s benefits and achieve the desired outcome. Follow our step-by-step guide above to know how to choose the test cases to automate and learn how to measure your result after automation.


Leave a Reply