We run regression because it makes us feel better — not because it adds any value
This blog title needs some context — it’s those regression cycles — either all automated, partially automated or manual, that you run all the time, yet reveal no bugs. You might come across flakey or broken automation checks that you need to fix. You start to ask yourself, what is the value in running these at all?
“But Melissa, it’s not about what you find. It is about giving us confidence.”
It feels a little ironic that it is making us feel confident by finding nothing. I would feel more confident if the checks actually found a problem!
It is those scenarios where instead of using the facts our emotions start to get in the way. But…what if we miss something *this* time? That gut wrenching feeling that we will get blamed for making a decision not to run it and something bad happens.
I guess this is where we can start talking about adding automated tests to the pipeline and monitor the test results. Hiding the fact that the effort involved in writing these tests might be a complete waste of time as they find absolutely nothing!
In a previous blog I wrote about risk based regression testing. This could be one solution to the problem. https://fishouthebox.medium.com/introducing-change-risk-based-regression-58969eb7bce9