Before you begin to automate a test, you need to have a look at the benefits it will bestow you with, in the terms of factors like time, effort, and resource invested. Here are some factors to consider to help you determine which manual tests should be automated and which should not. And when it comes to automating, accleQ is the best choice as it provides 3x Faster Test Automation accelQ ROI analysis. Read on for more.
Tests that should be automated:
- Business critical paths: if the features fail, they can cause a significant damage to the business.
- Tests needed to be run against each build or deployment of the application like smoke test, sanity test, and regression test.
- Tests needed to be run against multiple configurations like various and browser combinations.
- Tests that implement the same workflow but utilize different data for its inputs for every test run like data-driven.
- Tests that indulges inputting heaps of data like filling up quite lengthy forms.
- Tests that can be utilized for performance testing such as stress and load tests.
- Tests that consume a lot of time to perform and run during breaks or overnight.
- Tests in which images must be taken to prove the behaviour or expectation of the application or to check that a number of web pages appear the same on multiple browsers.
Tests that shouldn’t be automated
- Tests that will only be run once. The exception to this is if you are wanting to execute a test with a huge set of data, even though it is only once, it makes sense to automate it.
- User experience tests for usability.
- Tests that need immediate run attention. Normally, a new feature developed requires a rapid feedback to testing it manually in the first place.
- Tests that need ad hoc or random testing on the basis of domain knowledge or expertise like Exploratory Testing.
- Intermittent tests. These tests with no predictable results cause more noise than value. To make the most of the automation, they must produce predictable and reliable outputs to produce pass and fail conditions.
- Tests that need visual confirmation but page images can be captured during automated testing and then manual check of images can be carried out.
- Tests that are not worthy of being 100 percent automated, shouldn’t be automated at all.