Manual vs. Automated Testing. Where They Mess

Both manual and automated testing contribute to any project type, but particular cases can yield maximum benefit. Let’s find out what is the difference between manual and automation testing, where to apply a specific kind of testing, and how it can boost your project.

Cons of Manual Testing

Manual testing is performed by individuals, which implies the human factor. It makes the entire process less reliable, error-prone, and time-consuming. Some tasks can’t be performed manually, such as API testing. It's all because they require too many interactions or too frequent responses. So, a carefree attitude may lead to ominous consequences. 

Apart from that, scaling manual testing means scaling up human resources. If you want your fast test execution, you need to hire more people, which is quite costly. Add here the ability to execute (and maintain) the entire scope of up-to-date manual tests for short-term project iterations, and the usage of manual labor becomes questionable.

Pros of Manual Testing

QA Engineers often use manual testing to test visual aspects, such as interface or usability. Generalle, people do it better than a script. Testers act as end-users, providing beneficial feedback about the product. It is the right choice for small short-term projects due to the shorter execution period and instant review. So it is the best choice for startups in need of faster time-to-market. On the contrary, it requires too much time to create automated tests.

On top of that, our QA engineers provide more in-depth business analysis and specify project requirements in the early stages. It results in more transparent tasks assigned to developers and lets QA specialists start test cases before the development process.

Manual testing is resistant to minor changes. Although some effort is spent on keeping the test cases up to date. Automated tests, in its turn, may require updates even amid the smallest shifts. In these cases, human observation has more value than robust automation. 

Cons of Automated Testing

Very often, people think that automated testing is as a one-size-fits-all solution. The solution that removes all the project bottlenecks. But the experience shows that it doesn’t. One of the main QA principles is that one cannot automate everything. 

In real life, QA engineers provide automated testing for features that are:

1. Of high priority. Some functionality may be too significant to rely on manual testing and allow for human factors. 
2. Stable against changes. In the course of functionality changes, automated tests need updates as well. So sometimes, it is easier and faster to check something manually instead of rewriting tests over and over. 
3. Automation-friendly. Remember, you can not automate everything. Attempts to automate usability or exploratory testing are pointless. 
4. Risky. Some functionality may be too dangerous to “touch” it manually. In this case, a well-written algorithm is used. 
5. Complex. Some functionality requires complex frameworks and tools to perform automated testing. Not all engineers have enough expertise to do that.

So it may be resource-consuming to apply automated testing to small projects. The reason is that it is time-consuming and too intolerable towards changes. Not everyone is ready to sacrifice half of the development lifecycle on writing tests. 

What's more, come not into the favor of automated testing, it can't cover the entire project. Sometimes it covers 70%, but the remaining 30% have to be executed manually. 

Pros of Automated Testing

When do you choose automated testing over manual testing? In general, automated testing is used in large projects with a lot of repetitive and time-consuming tasks. Tests run by script do not tend to get “bored,” thus error-prone instead of the tests conducted by QA engineers. It is a good option if you need to save time and reduce human factors.

As an example, let’s look at one of Geomotiv’s case. We created Cypress-based E2E tests to ensure the high quality of a media project. With this approach, we managed to automate every user action scenario. Even the smallest features had their E2E tests with special mocks where the real app had to interact with external services. It provided a comprehensive feature review and ensured everything worked as intended. What was more, Cypress uses the BDD methodology. So even non-technical people can write full-fledged tests. 

With automated testing, you can reuse the existing tests, get faster feedback, and deliver accelerated results. Once you’ve written those scripts, they stay forever - pure reusability. 

And to have your tests well written, it's better to assign the task to an SDET. 

Who Is SDET

A Software Engineer in Test (SDET for short) is a developer, not a QA engineer. Its primary responsibility is to create software for testing, automated tests, and frameworks, tools, and test scenarios. As developers, they can understand the project architecture better. Thus, can create relevant test software. 

Their dual competence in development and QA contributes a lot to improving the testing process and ensuring its high quality. They work best with manual and automated testers, providing them with all the required software to perform their work effectively and competently. 

One should note that SDETs cannot take the entire QA department’s place. That's because they don’t test anything. Their task is to build software for testing, but not to use it. A single person cannot handle all activities related to testing. There are different specializations, and they should be applied together for various purposes.

Where Testing Shines

As you can see, automated testing works best when a task is routine or time-consuming. Simultaneously, it increases test coverage, speeding up the entire process. Provided that you’ve spent some time creating tests.

Manual testing is best suited for tasks that require human reasoning, creativity, and even intuition. Moreover, the manual approach provides quick feedback. It is not as time-consuming to set up as automation. “It’s better to test something manually than to test nothing automatically” - that’s a famous saying in the QA sphere.

What to Choose

If it is possible, choose both. More than 60% of automation experts won’t do any manual testing, while purely manual QAs might work on a project for too long and still pass over too many bugs. 

Both types of QA support each other and annihilate their disadvantages. Of course, a sound QA engineer should have both substantial manual and automated testing experience. However, it is difficult to find such an expert.

QA is a “project within a project” with its scenarios and designs. So, human interaction is needed anyway along with well-written automated tests. Buy nice or buy twice. Here, at Geomotiv, we have both types of experts for any kind of project. So, if you aren’t sure whom to have, ask us.