80% Automation 20% Manual

One of my students wanted to join Automation course. I generally ask candidates that ‘why they want to learn software testing?’

His answer was, ‘Ma’m there’s an equation which says 80% manual testing and 20% automation but I want to do exactly opposite.’ I immediately asked ‘have you seen any project released successfully with that equation?’ he nodded with an answer ‘No’.

I am neither against automation nor in favour. It depends on the context of your project. Automation would be creative and helpful if you would use it wisely. Success and failure of your efforts depends on your relation with testing tools.

Automation is nothing but another software development that means it is going to be buggy. So you need to have proper planning, gathering requirements and testing the testing suite.

Let me give you an example which would explain how automation can be useful. In hospitals, you may have noticed that there are giant complex machines, for e.g., scanner which scans patients’ body from head to toe; then there are X-ray machines and devices to check heart beats, pulse rate etc and many other devices.

• With the help of such machines doctor do heart, brain operations successfully.
• It helps doctor to take fast decision
• It monitors patience’s health performance
• Doctors can act immediately in the current situation
• And it makes doctor’s life easy!

But there are disadvantages as well:

• Machine may generate wrong report which directly impact human life
• Machine’s readings may not be accurate
• Malfunctioning can happen
• It means they also make doctor’s life more risk

Machines can be useful and risky as well. Risky because machines can be buggy and humans even can make mistakes.

Similarly automation in software testing would help you to speed up your work. Tester can find more time for doing better things in testing. Automation would give you comfort if tester would use tools wisely. Development team can get immediate feedback.

The main point is relying on automation may not be a right decision. Do you really think trusting more on machine is better than human in software testing?

It depends on the context of the project. Some project may not afford to spend time and money on automation they would find this as luxury. Some project may need it because task can be extremely critical where humans can’t reach and perform them. For example manually (using hands) digging a huge coal mine is impossible for human.

Aeroplane is a machine but you need human to fly without their skills things can’t be effective and may not work.

We need humans because they can observe, take decisions, notice, think, feel, learn, situational awareness, communicate, tell the stories, present, develop new skills, investigate, collect information, analyze, infer and take charge of situation.

I feel that software testing has lot of aeroplanes but we need skilled pilots who are expert, trained and intellectual to handle them in various situations.