Regression testing is an integral part of the testing ritual.

Regression testing seeks to uncover new bugs in existing builds after changes. The process might be difficult at times to keep up with, but the benefits it brings are significant. Without conducting regression tests, your system will suffer from defects and become unreliable.

According to IBIMA Publishing, only 7% out of 45 research papers focused on regression testing.

fig 2
This shows that the topic of regression testing is still undermined by testers and researchers. It’s not always addressed by the amount it deserves.

Here are how regression testing will contribute to your product’s quality:

1. Maintainability

When you push your code to production environment, it becomes necessary to go back and check whether everything works. You can’t assume your code will still work the same way it did on your local environment. Something changed, and there will always be consequences.

It’s best to perform regression tests every time the product changes. It will be much easier to detect new and unexpected bugs within the system. Thus, making sure that the system still meets the product’s requirements. That leads us to our second point.

2. Reduce defects


When you try to patch a bug, you introduce several new bugs. In some cases, your code structure may change after one variable changed.

To reduce defects effectively, you need to know bug fixes and understand how they affect the system. Look for code churn and dependencies.

You should ask yourself these three questions:

  1. Do I have this bug somewhere else?
  2. What bugs may occur after I fix this?
  3. How do I prevent bugs like this?

Consider the testers’ holy grail – “automation”. Automation can be built-in unit tests that runs after each new build to make sure you did not introduce new bugs. Regression testing would require a tester to go back and test one part repeatedly. Automation will help to reduce a significant amount of work. Check out Testlio Testing Tools Store. There’s a variety of tools for you to choose from and install.

But you have to be careful when it comes to automation. Everything is not as easy as it seems. Excessive use of automation might make you become more dependent on the tools than your own ability. Problems can be right under your nose and you won’t see it. Why? You were only looking at the obvious or the results from your automation script. It is important to make sure that your tests cover all the edge cases, they are tricky. A combination between manual testing and automation would be ideal.

3. You developer may not be so informative

One of a tester’s biggest struggles.

Sometimes, developers may make minor changes here and there. However, they might not inform you about these new changes. They either forgot about them, or they thought that the changes didn’t matter.

However, in most cases, there will be unintentional errors, or what we would call “bug creep”. Did they use that string length anywhere else? Did anything else also change when they introduced that new condition?

Always trust your own gut. Be skeptical. Be loyal to your testing approach from the beginning until the end. Test back and forth. Regression testing is not an option, it’s a requirement.

4. Ensure robustness

To make sure everything else will still be working as intended even when one function crashes, a lot of regression testing has to be done. Dependent variables have to be mapped out and tested thoroughly.

The best way to increase robustness is to reduce dependencies.

Many platforms are starting to develop applications as a suite of small services instead of building everything in one place.

Traditionally, when everything is built from one place, one change can affect everything. It’s not effective and efficient. If one function crashes, you have to do regression testing on everything else and this will take up a considerable amount of time.

Focus on doing one thing and do it well. When something went wrong, you will immediately know where the bugs came from. There will be much less things to test since you don’t have as many dependencies.

You should always understand the core features of your products while doing any kind of testing. Know what is necessary to keep and what isn’t. Spending too much time testing and writing up a bug report on problems that don’t matter as much will waste everybody’s time. If there’s a bug in a small function and other important functions aren’t affected, then it’s okay to leave it out sometimes.


5. Save your time and money in the long run

Regression testing is an integral part of performance tuning. You need absolute confidence in your system before you introduce your product to the market.

You don’t want a buggy product. You don’t want to ship code you are not confident about. A detailed test suite is recommended.

If you introduce a buggy product, you’ll spend a lot of money and time fixing it. You’ll lose users that you normally wouldn’t. You can avoid this by taking care of your product a little more. It’s generally hard to convince users to give your product another try.


Regression testing is not an option, it’s a requirement. Regression testing discovers new defects coming from product changes thus helping you to ensure the quality of your own product. Running regression tests will:

  1. Help to maintain the product
  2. Reduce defects
  3. Prevent bug creeps from lack of information and communication
  4. Ensure robustness
  5. Save your time and money in the long run


What are your opinions on regression testing? What do you think are the best approaches? Did this post apply to your situations? We’ll love to hear your comments below, or you can tweet me at @anhthu_le179.