Especially in quality business, we’re very much used to control processes via product characteristics when setting up the process, then let the process run under no control unless we sense some problem, ultimately checking the process again at the end of the production batch, via product characteristics once more, before delivery to customer.  This is the essence of process – control and SPC.

This preceding is typical both for tangible products and for services.

THE PROCESS CHALLENGE
When asking Statisticians about sampling plans, we most probably will be answered in terms of raw materials and of finished products, and much more rarely in terms of sampling plans implemented on a product being made.

I’m italian, I love spaghetti, and they have to be boiled the right way.  It’s a process.  After six to seven minutes in salted boiling water I taste spaghetti, then every thirty seconds or so, until I feel their toughness is right. And when we say that a steak is well done, we mean much more than what’s served us on the dish.

Let’s think of a statistical X-R chart, or any run charT.  You would pick up a pre-established number of widgets.  You would measure some pre-established characteristics and you’ll plot them on a nice sheet of paper or on a touch screen.

But you’ll never know what happened to your product’s characteristics from the latest sample to this one.  What I mean is that process measures have to be dynamic and based on process – not end product – characteristics.

I may drive my car from A to B, and keep the average speed within law’s limits; but, in any moment, I might exceed them, and therefore be liable to be fined.

THE RISK CHALLENGE
We’re facing the same risks when dealing with risk.

Let’s think of risks as processes and represent them with lines, starting from point A and finishing in point B.  We typically only consider Points A and B, that is, the beginning and the end of a geometrical segment, but seldom consider the line between the two extremes, that is, the process itself, as it is.

The performance indicators requirements put a further bias on this approach.

Of course, some approach to continual process monitoring may reveal itself costly and unproductive; yet, there’s all evidence that this is the very core of risk-oriented process management: that is, to analyze processes and prioritize their controls while they run, in order to reduce severity and occurrence of their predictably negative consequences.

In-progress process metrics monitoring (IPPMM) seems still to be a far away discipline, despite the many and much charted process parameters.  We still have to clarify to ourselves what a process is and what isn’t, especially when dealing with risk.

We might end up in managing disaster because we were too easy going when managing risk and this would obviously weigh on our shoulders.

Due to obsession with time, we’re more and more often using timing devices; but we don’t know neither we validated the time sampling plans we use: do we really know how long does it take us to awaken, shower, breakfast, check our luggage, drive to the airport for an intercontinental flight?

What about if our car’s engine won’t catch up? Or our neighbor’s tractor went astray on the road, just in front of us?

Beginning and end controls belong to the realm of predictability, while in-progress process controls are what they are; they mean themselves and the process’s actual reality – which is often unpredictable.