Industry Wisdom Test Planning & Management

Automation is Confirmatory, Manual Testing is more Exploratory?

Once upon a time,

Me: Let’s run Automation for regression tests.

QA Lead: Yeah! But please ask team to do manual tests as well.

Me: What! But we are running automation suite?

QA Lead: Automation is Confirmatory, Manual Testing is more Exploratory.

Automation is mostly employed for regression, i.e. to gauge the impact of changes on ‘already-working’ functionality. Impact as in – to confirm already-working functionality is still working. The same Manual test cases are automated and executed. Automation is for confirmation.

What if you want to find issues/defects? Yeah! Manual testing is the first approach. After all, end-user is not a scripted machine. Simulating end-user behavior, testers explore the application with some alternate flows & on-the-fly data. Thus identifying some unknown (not yet caught) defects. In that sense, Manual tests are more ‘real’, more closer to the end-user behavior, i.e. more exploratory!

Got an opinion? Please comment and let’s discuss!

………………………………………………………………..

Ramdas Krishna Baliga | Staff Software Engineer in Test at Intuit

I slightly differ here. For me Regression means testing / regressing to ensure new code changes did not break working code. Manual / Automation are means to run it.  There are cases u cannot automate due to various constraints and those will be covered by manual effort. Return on Effort is used to decide the approach. I am still in favor of more automation provided it cuts effort (and does not add overhead of monitoring false failures).  But in most scenarios I have seen automation is written only after feature is stable 🙃 to guard against future failures.

Manual vs Automation Testing

NIRMAL.G | QA Consultant

Exploratory Testing avoids the pitfalls of pesticide paradox. Automation Testing confirms that the functional flow is good.

………………………………………………………………..

Vishal Gondalia | QA Team Lead at Paperchase Software Solutions

Automation is just for check that the functionality is working proper or not, but while in manual you can able to perform monkey testing for better output which cannot be performed in automation. It’s only for one defined flow testing. From my view, manual is best practice, yes but while performing regression and sanity testing then you can perform automation testing which is helpful.

………………………………………………………………..

Rajeev Vaidya | Software QA veteran. Aspiring data scientist.

Automation testing should be used to automate the repetitive tasks of manual testers.  If used wisely, automation can help in unlocking the value of manual testers. Once he is free from repetitive tasks, his end-to-end knowledge about the applications under test, demand for quality, ability to interact with the stakeholders can be leveraged.

………………………………………………………………..

Rahul Anikhindi | Consultant Specialist (Test Lead) at HSBC Technology India, ISTQB, TOSCA Specialist L1/L2

First of all both methods of testing has different objectives to accomplish. Manual is performed when any new changes are put on the software.  Automation is not useful at this stage and should be used for regression / when build is stable and to be run independently to run such scenarios with multiple combinations of test data to save manual efforts. Happy testing 🙂

………………………………………………………………..

You cannot be all time dependent on automation testing, manual testing is always required.

Exactly… Manual testing is must to unsure new functionality is working fine or not. Automations is only limited to regression and sanity test suits.

Testing cannot be automated. What can be done is automated checking. We only can put automated checkpoints.

Manual testing is mandatory to perform while Automation testing depends on feasibility in the project.

………………………………………………………………..

Shashank Tiwary | Sr. Software Developer

Automations​ testing is a kind of test process and it is not a complete solution.  In general I apply this for continuous integration. It confirms the merge feature is not breaking anything. There can be another use of this.

………………………………………………………………..

Priyanka Babbar | Business Analyst at Telstra Health

There has to be the right balance of both, Quality cannot be guaranteed with only Automation testing. Human mind to think from user perspective cannot be scripted to its full potential. 

………………………………………………………………..

Michael “Fritz” Fritzius | The Test Automation Guy

If you use manual testing solely for finding defects, automation is useless. But if you automate every manual test, the test suite will crush under its own weight.

The best automated tests I’ve seen are the ones that, when they fail, show what part of the system needs a human to poke around on more. There’s just enough of them to say, “Hey, something’s wonky over here.”

I call it “Tripwire Automation”.

The reason why this is good is that those automated tests stay viable longer. There also shouldn’t be nearly as many to maintain, and having fewer tests leads to fewer false positives.

………………………………………………………………..

Automation depends on Scope and feasibility of project. At the Time of Major/Minor release once have to go Manually.

Manual + automation is the way to go…Create test data/setup using automation but test manually.

Automation is a great solution for reducing regression cycle time but can never replace exploratory and ad-hoc cycles where most complex issues are found.

Looking for a Job Change in QA/Testing technology? Get Regular Job notifications @ WhatsApp!

Leave a Reply

Your email address will not be published. Required fields are marked *