The life of a QA Tester will be considered incomplete if the terms ‘Smoke Testing’, ‘Sanity Testing’ and ‘Regression Testing’ are not infused into it. Though these are some regularly used terms, there are some common misconceptions around them too.
Before we begin into the details of smoke, sanity and regression testing and what they actually mean, lets go through some common misconceptions and myths around them:
- Smoke Testing and Sanity Testing are the same and can be used interchangeably : This is not true and should never be done, why? we shall discuss that in this Article.
- Sanity Testing is equivalent of acceptance testing: Acceptance Testing is done to ensure that the build meets all the requirements that the client specified before the release while Sanity Testing is done to ensure the product is sane, rational for more detailed testing. These are not and should not be used interchangeably.
- If I am doing Smoke Testing, I can skip sanity testing: Smoke testing is a very high level testing, as in, this testing is done to ensure this build is suitable to begin any other type of testing. For example, this could be just a test to ensure the build installs and then the Sanity Testing can begin. Though sometimes, the smoke test cases are run together with Sanity Test cases, these should not be confused to be the one and the same thing.
- Regression testing is not related to Smoke Testing or Sanity Testing: In Theory, Sanity Testing is a subset of regression testing. Some high priority regression test cases constitute sanity test case usually.
- If I intend to run the full regression test suite – I don’t need to run any smoke test or sanity test: While this may be true in some cases as Sanity Testing is anyways a subset of regression testing, it is still advisable to execute the sanity test cases first and then follow them with the rest of the regression test cases.
In this article, we will try to clear above confusions for once and for all.
In case, you wanted to read more about Manual Testing, do refer to the link here.
Note: Because we will be using the term ‘Software Build’ many times in the article, lets define it here. ‘Software Build’ is a process of converting source code, into a user application, after multiple revisions and code changes and Software Build creation involves multiple processes like, “Version Control, Code Quality & Compilation” and software build is also the result of this building process. In this article, a build will be referred to as a testable version of the software.
Smoke testing is an approach which is usually carried out during the initial development stages of the Software Development Life Cycle(SDLC) to make sure that the core functionalities of a program are working fine without any issues. It is executed before any detailed functional tests are done on the software.
The main intent of smoke testing is not to perform deep testing but to verify that the core or main functionalities of the program or the software are working fine. Smoke testing aims to reject a badly broken build in the initial stage so that the testing team does not waste time in installing & testing the software application.
Smoke testing is also called as Build Verification Test.
Let’s see a simple example where you are given an email application to test. The important functions would be logging in to the email application, composing an email and sending it, right? And, in case the email is not sent, does it make any sense to test other functionalities like drafts, deleted messages, archives, etc? This means you will have to drop the build without further validation. This is called Smoke testing.
The main focus of smoke testing is to test the critical areas and not the complete application.
When to perform Smoke Testing
- When developers provide a fresh build to the QA team. A fresh build here means when the build has new changes made by the developers.
- When a new module is added to the existing functionality.
Automation & Smoke Testing:
Usually this is the type of testing that is executed before actual automation test cases can run. For organisation that have continuous testing built in, smoke testing is equivalent to successful installation of the build for running test cases or execution of the first test case. So, this is not a type of testing that is deliberately automated but if test automation is put into place, the test automation can only run successfully once the software has passed smoke testing. Or otherwise, the first test case that executes might fail.
Sanity testing is a kind of testing performed to check whether a software product is working correctly when a new module or functionality gets implemented to an existing product. Sanity testing is a software testing technique which does a quick evaluation of the quality of the software release to determine whether it is eligible for further rounds of testing or not.
Sanity testing is usually performed after receiving a fairly stable software build or sometimes when a software build might have undergone minor changes in the code or functionality. It decides if end to end testing of a software product shall be carried out further or not.Sanity testing is also a Surface Level Testing which helps in deciding if the software build is good enough to pass it to the next level of testing.
Why perform Sanity Testing
- To verify and validate the conformity of newly added functionalities and features in existing code.
- To ensure that the introduced changes do not affect other existing functionalities of the product.
- To decide further testing can be carried forward or not.
When to perform Sanity Testing
- Build is received after many regressions or if there is a minor change in the code.
- The build is received after bug fixing.
- Just before the deployment on production.
Automation & Sanity Testing:
Considering, Sanity Testing is considered as a subset of regression testing, these are the test cases that can be automated. A recommended approach is to execute these test cases before running the complete regression test suite. The benefit is that if there are any errors in the sanity test cases, then errors can be reported sooner rather than later.
Regression testing is the verification of “bug fixes or any changes in the requirement” and making sure they are not affecting other functionalities of the application. Regression testing is effective on automation and usually performed after some modifications have been made in the software build after requirement changes or bug fixes.
Once Sanity testing of the changed functionality is completed, all the impacted features of the application require complete testing. This is called regression testing.
Whenever bug fixes are done in the existing software, some test scenarios need to be executed, to verify the bug fixes. In addition to these, the QA team also has to check the impacted areas, based on the code changes. In regression testing, all those test scenarios will have to be executed, to take care of related functionalities.
When to perform Regression Testing
- After Code modification according to the required changes
- After some new features are added to the application
- After some bug fixes are incorporated into the build
Automation & Regression Testing:
Regression test cases are actually the ideal test cases for automaton. Usually, when an organization starts automation, these are the test cases that are automated first. If regression testing is an activity that is taking a lot of time for your testers and the same test cases are repeated multiple times then it is time that you start thinking of automation too.
If you are looking for a tool that can help you get started on your automation journey, then you also need to ensure that you choose the right tool. A tool that can also provide you ROI on the efforts invested. We have the guide that can help you there:
Differences Between Smoke vs Sanity vs Regression Testing
|Smoke Testing||Sanity Testing||Regression Testing|
|Performed on initial builds||Performed on stable builds||Performed on stable builds|
|To test the stability of new build||To test the stability of new functionality or code changes in the existing build||To test the functionality of all affected areas after new functionality/code changes in the existing build|
|Covers end to end basic functionalities||Covers certain modules, in which code changes have been made||Covers detailed testing targeting all the affected areas after new functionalities are added|
|Executed by testers & sometimes also by developers||Executed by testers||Executed by testers, mostly via automation|
|A part of basic testing||A part of regression testing||Regression Testing is a super set of Smoke and Sanity Testing|
|Done usually every time there is a new build||Planned when there is not enough time for in-depth testing||Usually performed, when testers have enough time|
- Smoke and Sanity testing help the QA team save time by quickly testing to make sure if an application is working properly or not. Also, it ensures that the product is eligible for further testing. Whereas Regression testing helps enhance the confidence about the software quality after a particular change. Especially, that the code changes are not affecting related areas.
- Smoke Testing is done by both the dev team or by the QA team and can be taken as a subset of rigorous testing. Whereas both Sanity & Regression testing are done only by the QA team. Also, Sanity testing can be considered as a subset of acceptance testing.
- Smoke testing is executed at the initial stage of SDLC, to check the core functionalities of an application. Whereas Sanity & Regression testing are done at the final stage of SDLC, to check the main functionalities of an application.
- As per the requirement of testing & time availability, the QA team may have to execute Sanity, Smoke & Regression tests on their software build. In such cases, Smoke tests are executed first, followed by Sanity Testing & then based on time availability regression testing is planned.
In practice, all QA teams need to do Smoke, Sanity and Regression testing. All of these testing types have a pre-defined number of test cases that get executed multiple times. This repetitive execution also makes them an ideal candidate for test automation. When looking for automation, you are recommended to use a tool that provides you ROI on automation from the initial stages. Testsigma is one such tool.
Choose a tool that lets you automate from day 1