It is a well-known fact that every person in the world makes mistakes. That is why, today, we will talk about the most common errors made by manual testers while creating test cases.
1. Maximum Simplification of Test Cases
Sometimes testers can simplify test cases to such a technical condition so they cannot be conducted according to an algorithm.
Here’s an illustrative example of such a case. After choosing a user in a table, we’ll see a context menu with the following parameter: Edit, Create, Delete, and Copy. One needs to ensure that all the available functions are working properly.
It will be extremely difficult to automate this process. First of all, if a tester clicks on Delete, he/she will close the whole context menu. There will be a need to reopen it. And such scenarios aren’t described anywhere.
Second of all, the user will be deleted. To test the performance of other functions, one needs another user’s account. Once again, these scenarios aren’t mentioned anywhere in specifications.
2. One Step Contains Forwarding to the Content of Another Step
It is not recommended, for example, in step #10, to create a test: “One should repeat steps #8 and #9 for the user of the admin group”.
Such a test has to be absent at all, or it should be in a separate step. Such forwarding can increase the time of test release in 2 or 3 times.
3. Branching Within One Test Case
Here’s a simple example. Go to the user page. Check if a table with the email, name, and type columns is displayed. If there’re no users, ensure that the table is present.
There is no need to do these things. A test must have one way which is defined by some pre-conditions.
4. An Attempt to Automate Irrelevant Test Cases
QA specialists have to check test cases before proceeding with the automation stage. It is worth remembering that automated tests are developed much slower comparing with test cases. Hence, instead of giving about 30 tests to automation, one should divide them into small parts (4-5 tests) and give them one at a time. In this case, there will be no need to spend much time finding potential reasons for incorrect test performance.
So What Should We Do in Such a Situation?
Here we have simple but effective recommendations:
- You should always write tests inside the corporate test tracking system;
- Create them in details;
- Try to avoid repeating tests and test steps;
- Describe all the possible auxiliary data for testing;
- Try to divide test cases in such a way so automation starts with the most important tests;
- Don’t combine a lot of tests (especially complex ones) into one test.
Obviously, one may wonder: “Why is it necessary to adapt created test cases to automated tests and not the reverse?”
The answer is really simple: automation is a very thoughtful and expensive process. The speed and quality of the automation process largely depend on which specific test cases will be given to the automation engineer.
Unfortunately, a lot of test cases created by manual testers cannot be automated. Specialists have to redo and rewrite them. And automated QA doesn’t always have time and desire for such things.
In other words, the success of test results depends on manual testers. If automated QA receives high-quality test cases in the first place, there will be good autotests at the end. This means that the software will be of high quality and in-demand among end-users.