Predictions You Can Count On --Dickeson

We know that we cannot predict the national deficit, the stock market, our sales, the weather for more than a few hours, the fingerprints or blood genomes of an individual, and almost any sequence you can name.

Oh yes, poetically, you can predict that the sun will rise in the east tomorrow. But, actually, what precise second and where is east? The world is not linear and never has been. For centuries we assumed it was for the sake of rounding-off conveniences. The computer brought a screeching halt to that idea in the last half of the 20th century according to scientists in all disciplines.

With computers, we stopped rounding Pi to 3.1416 and began carrying it out to thousands, or millions of decimal points. That's what dumb computers get paid to do in micro-fractions of a second. The world of convenience rounding gave way to a surprising non-linear world where straight lines don't exist.

Variation is the rule of life. Some called this fractal reality; others termed it chaos. Very small changes in starting or operating conditions have, we discovered, profound impacts on output. This is the law of fractal life.

Lessons From the Past

We learned this, but we didn't adapt with the knowledge that came to us. Job cost is a linear system. It starts with linear budgets that are impossibilities in our fractal world. We learned this in the late '80s and early '90s.

We vainly tried to predict the number of hours we could charge to jobs in every work center a year in advance. Impossible budgets were divided by foolish chargeable hours to arrive at a Budgeted Hour Rate. Topping that, many of us added contrived allocations of sales and administrative costs to the budgeted rates! We took the linear impossibility to a ridiculous extreme.

How did we use this rate? We then multiplied it by "standard" hours: predicted hours for pricing and loading and actual hours for post-mortem analysis. But there are Five Feuding Families in our printing process: a) people, b) machines, c) materials, d) environment and e) sequence. Small variations within, and among, these Five Families always cause non-linear variation in hourly outcome dollar values.

Standards are linear predictions, while actual hours are linear outcomes. Then we impose on these the Voice of the Customer, seeking linear demands of quality attributes and delivery schedules—budgets or goals—that we can only incompletely fulfill.

Managing is predicting. This we know. We must be able to predict the results that will follow from our actions in order to run a business. But now we know that we can do so only imprecisely because there's some master scale of rounding and interrelationships—input variance—that confounds precise attempts. We have to live in this imprecise, non-linear world. Somehow we must go with the flow and do the best we can. We can't predict with precision anymore so we "optimize" not "maximize."

With humans of variable skills, machines of a variety of capabilities in variable conditions of repair, materials in a range of conditions, environment shifting between humid and dry, hot and cold, dusty and clear, clean and dirty, and prior jobs requiring settings, undoing and re-doing. . .we cope. We perform in an ever-changing fractal world where slight changes in variability-scale or rounding have profound effects on outcome.

BHRs (Budgeted Hourly Rates) and production standards—Job Costs—are products of an old linear world that no longer exists and never really did. We did the best we could with what we had to work with at the time. It was our own version of reality because we believed in it. It became true because we said it was true, but in our hearts we knew it wasn't so.

Now it's all gone with the wind. Never really was. The thoughtful ones sensed that BHRs and standards were foolish and began to abandon them while still paying lip service to the old ways to avoid controversy. Besides, what else was there but gut-feel and luck?

We'd spent thousands of dollars on Management Information Systems that were premised on job costs—flawed linear assumptions. Who's ready to admit that it's all outmoded? There were, and are, those who still believe BHRs and standards are true despite all the evidence to the contrary. Do you?

In that old linear world we believed that jobs caused costs. That was the implicit assumption. Increase efficiencies—cut job costs—and enhanced profits follow. Not true. Actual expenses don't change. Job "costs" may change, but all you did was move some expenses to marketing and sales to cover with increased revenues. But that's another story.

Now we live in Sigma World, a non-linear world—a fractal world—the world where chaos prevails. Sigma is a calculation of one standard deviation from the average. We now believe that within three Sigmas on either side of the average, 99+ percent of all normal variation will occur. That's a stable range we can live with for predicting, for managing our business. If incidents occur outside those limits they're attributed to "special" causes that must be explained and eliminated to establish predictable stability and continuously improved performance.

In a nutshell, that's my understanding of the current non-linear reality. Small changes in conditions cause profound changes in output. That's normal, non-linear, variation. It's a given for any system. We predict what will happen within that range and we live with it. If it's within range we shut up. If we don't like the range of prediction, then we must make substantial changes in the system of Five Feuding Families of Printing.

The easiest, quickest way to find the range of truth is to use the shortcuts developed by Walter Shewhart more than 60 years ago, preached by his student W. Edwards Deming, and now taught by Donald Wheeler in his "Understanding Variation, the Key to Managing Chaos." The path to prediction is through the use of process behavior charting (called XmR). It's now offered as a part of CRC Information Systems' software, a printing MIS application. It's also in use by a large, multi-plant printer in this and other countries. You can get a package add-in for Excel at www.quantum-i.com to use on your own files.

Be warned, however. If you're still using BHRs and Production Standards from the ancient age of linear calculating, you won't like the new results. That's your problem. The world you now live in is fractal, not linear. Change or die. You choose.

You can only predict within a stable range. You can only manage what you can predict.

—Roger V. Dickeson

About the Author

Roger Dickeson is a printing productivity consultant based in Tucson, AZ. He can be reached via e-mail: rogervd2@cox.net.