Explaining the importance of testing

Sometimes it may be OK to not test, but first you need to ask what is the consequence of finding an issue.

If a change doesn't affect the customer's functionality such as it is limited to your own internal data collection on the SW performance, then it might be acceptable to not test, but then you also don't know if your internal data is any good. Another example is if the functionality changed isn't needed now but we are inserting it early to gain practical experience, and it is isolated from other systems

Otherwise, the change may adversely affect the intended function, or inadvertently impact functions either within the changed module, or other modules with different owners who interface with the prime group.

In a safety intensive industry, if someone wishes to avoid functional or regression testing, then they must justify their rationale. That is treated on a case by case basis and reviewed with cross-functional team of equipment owners to ensure potential adverse affects were not overlooked.

Good point, we can ask test resisters to define the consequences when something goes wrong, that may get them thinking differently. Usually, the consequence is "we did it wrong" and "they have been jilted" by our failure. So you can see, this gets pretty touchy. So we end up in a bit of a tug of war, with customers saying, just put the data in Prod and no we don't want to test, but if anything goes wrong, it's because you failed. Not that we as a wider team failed to take the few minutes that is required to run a quick test.

Yes, unfortunately I know that situation well. "We don't have time to do it right the first time, but we always find time to fix the defects later."

I would personally try to explain the Cost of Poor Quality (COPQ) to the team. Most of my references are very technical, so I try to explain it more layman's terms. Some sources put the typical cost of waste as 20% of total revenue, but that can range from less than 1% to over 40% depending on the error rate. This article is a bit dry for most people, but might provide ideas:

https://www.isixsigma.com/implementation/f...-failure-costs/

Is this not a risk-benefit consideration? Benefit of testing versus risk of not. If so, maybe it is not a one-answer question. If the cost of testing is $50 and the cost of failure is $1000 with a 10% probability thus $100 exposure - then spend the $50 and test. If the exposure is equal to or less than the cost of testing, don't test.