Do you have any test automation for your project? If so does your management think its a magic robot that can pretty much do anything and that too on its own?
We have been asked to automate our project. It has to test everything, catch all defects and need no manual over site time. We try to explain that creating and maintaining test scripts is really important and project will have to scope in time for this every release.
A fellow blogger in the post Automation: Oh! What A Lovely Burden! talks about automation being a lovely burden.
What are some automation myths and how to manage them?
- Automation will test everything!
- Automating every single test is not a good investment. There are some features of the product that might never be used or used very little. Spending time on automating them may not give good return on investment.
- The more complex a test, the harder it is to automate it and the harder it is to maintain the test. If it takes a hour to manually test it during regression then just manually test it.
- Automation will catch all defects!
- Again this is a misconception. If something changes in areas that are already automated the automation test will catch it. If its in features that automation does not cover then we wont be catching it via automation. For example an automation test might check a box and continue with the test. A defect occurs when the box is checked and then unchecked. This defect was caught my a customer because its not a standard step that customers follow. It was a one off scenario.
- Automation will not catch defects in new features since they have not been automated yet.
- Once automation is complete no more time will be needed for maintaining those tests.
- Any change in future releases can impact automated tests. Time has to be spent in maintaining these scripts. It wont take as much time as it takes when creating new tests but its still time that testers have to spend and this time has to be planned into the project plan.
Yes automation is great but its not a robot and it wont do everything we would like it to do. It makes manual testing effort easier and also gives time to explore and test other features that are new or need our attention.