A lot of people that I've spoken with recently in non-regulated or loosely regulated industries have been dumping their calibration programs as a "cost-cutting" measure. I think we've all seen the pattern over the last 15 or so years of businesses trying to do more with less. But, at what risk?

Risk is everywhere. We all deal with it daily, so to the level of becoming agoraphobics that refuse to leave the "safe place" in their own lives. I think we are all comfortable with some degree of calculated risk in our lives. We get in the car each morning knowing we are transporting 10 or so gallons of highly combustible liquid a few feet from us, regularly at speeds that weren't even possible a century ago, and we call it our morning commute.
I think there is an inherent level of human nature that says, "it won't happen to me." Honestly, I think that is a good thing, without that gene the human race would have died out eons ago. But, we all know the situations where we take it a bit too far. Usually, we are talking about situations where the risks are extremely high, but the odds are extremely low. When we do head out on our morning commute any reasonably responsible person has insurance. Why? Because we know what can happen, and we know the cost of the "could happen."

Look at the cost of insurance. We all complain about it, but in all honesty, hundreds to thousands of dollars a year to protect someone from potentially millions of dollars in liability is cheap. And those of us who get to be considered "low risk," there's that word again, get away less expensively than those of us that have had an issue or six, or seven. 

In the regulated business world, calibration is a "shall" and not a "should." In the non-regulated world, it becomes a should. Even under the guises of 9001:2015, it is a should, it comes down to risk aversion, as a strategy therein.
Consider the situation of a company that exists in a consumer goods industry. They make the grandiose hard-wood doors, the kind you find on mansions and manors with titles that start with the estate of Lord such and such. They never calibrate anything, they never have. They decide to purchase new tape measures, to standardize. They buy their tape measure from a company that doesn't calibrate. The company never saw a need for it either. The machine that prints the numbers on the tape measures malfunctions one day and it runs out of sync with the device that feeds the tape through the system. Somebody notices the issue, so they fix it, but they ship the tape measures anyway. They check the first 12 inches of the possibly incorrect tapes, against good ones and the defect is hardly noticeable. They should be close enough, right. Back, to our elegant architectural door company. It turns out all of their tape measures came from the batch made that day, and when they make a door that should be ten feet tall, it is about a ΒΌ inch too short. How many doors do they ship out before they realize that they have a problem? How much is it going to cost to replace all of those doors? They could have sent out a single one of those tape measures for calibration at the cost of $40 - $60 depending on the calibration company. 

I describe this scenario because I've seen it more than once. The door company comes to me because they are in trouble. A lawn company seeks me out because their flow meter was missing a paddle on the impeller of their flow meter and killed someone's lawn with too much fertilizer. A tire installer has me check their torque wrench after a wheel came off a car. 
I'm not going to tell you to calibrate "everything." I don't believe that you should, but I do think you need to apply some common sense to measurement traceability and consider your risk tolerance. For me, the scariest part is how few people are aware of measurement traceability. I'll have a conversation with the general public, and they will mention measurement of some sort, and I'll ask, "how do you know it's right?"