Impact of Automation Testing and Manual Testing on SDLC

Introduction

Testing being exhaustive process, automation testing has a lasting impact on software development life cycle. As the automation testing is still evolving, it has lot of scope for enhancement of testing processes are carried out in SDLC. On one side manual testing is contributing in understanding the problem domain on the other side automation testing is contributing in execution efficiency and testing with variation of data-set with no margin for human error.

Importance of Specification Testing

Much of the manual effort is required in coming up with requirement specification, automation can derive tests from the specification but reverse of this process is not possible. As the whole development process is dependent on SRS. Correctness of SRS needs to be checked manually, similarly, SRS needs to be checked for ambiguity and consistency and completeness. There is no substitute for manual testing in this area.

Regression Testing

Checking that what has been corrected in software still works after correction, this phenomenon leads to regression testing. Automated testing tools are required for this type of testing. Manual testing is not appropriate for handling regression testing, as most of the tests are repetitive and automation testing can handle it better and are cost effective methods of testing as opposed to manual regression testing. Generally, organizations automate around 60% of regression tests from whole of regression suit which can be set as good benchmark.

Random Testing

Testing performed manually by using deep knowledge of software under test has outperformed the routine automated tests or specification based tests. The time saved by automated tests brings a lot of scope for random testing where testers uses their creativity and intuition for performing testing tasks. In this type of testing, manual testing scores high and yields greater defects than traditional specification based testing.

Skills

In recent days, software products and application are adopting AI and machine learning based technologies. With the evolution of AI and Machine Learning in the field of test automation, the skill of testers needs to be diversified. Skills such as AI testing, neural network optimization, business intelligence, algorithm analysis, understanding data flow and models along with automation frameworks and tools. It is predicted that 70% of the repetitive testing in future will be supported by AI based testing and 30% by monitoring and devising new strategies. While skill requirement for manual testing would be for usability testing, expert testing and testing from the perspective of look and fell.

Risks

There are certain risks involved while using the automation testing in SDLC as follows:

1. Project budget
2. Change in process due to introduction of automation
3. Unrealistic expectation from automation tools
4. Maintenance of automated tests after changes in software under development



There are certain risks involved while using the manual testing in SDLC as follows:

1. Increased risk of failure
2. Increased time and resource requirements
3. Everything cannot be tested manually

Conclusion

To conclude, automation testing and manual testing are two sides of same coin. While automation testing benefits execution efficiency, reliability and consistency, manual testing offers expert testing, using domain knowledge of tester and testing the requirements specification. Nevertheless, with rapid adoption and evolution of automation testing diversified skills are required to carryout automated testing tasks especially working with AI and machine learning based applications. Both manual and automation testing has certain risks when mitigated proactively, tangible benefits can be obtained.