An Ogre’s Guide to Quality Systems

Found this old gem in my archives:

MISSING  OCTAVE  INSIGHTS,  INC.
Newsletter 1/24/07
© 2007, All Rights Reserved
http://www.MissingOctave.com

An Ogre’s Guide to Quality Systems

To misquote a famous green ogre, Quality Systems “are like onions.  They have layers.”

We’ve all seen management try to force-fit a one-size-fits-all quality system on a business or product, usually because they want to get some certification or registration.  The result is usually fodder for a Dilbert comic strip.  But are each of us guilty of the same thing?  Do we each have a favorite approach to quality that we use blindly – one tool for every job?

Let’s peel back the onion and try not to cry at what we see.  Each and every product or project will reside at a different layer.  Our job is to find it, and, if it makes sense, to push it deeper.

Each of the sections below represents a layer of the quality onion.  Each must be broken through, in this order, before the next layer can be seen through the blood, sweat and tears.  You can’t be worried about Statistical Process Control when the machine is broken and the operators don’t use it anyway!

Layer 1 – “It’s Eleven PM – Do You Know What Your Children Are Doing?”

It takes a firm green hand to lead an effort.  But act like an Ogre, with too firm a hand, and the resistance goes underground.  Treat the operators like children and they’re sure to spend their energy trying not to get caught.

Operators are not children, but neither are they robots.  A human (unlike an ogre) worries about their place in the world and must feed their pride by exercising their own judgment.  The more experienced an operator is, the more likely they are to a) not read the instructions because they think they already know the drill, or b) do it their own way.  And of course each operator must show their own value by doing it THEIR own way.

The only way to lead a group of individuals to uniformity is to engage them in the effort.  If you’re an Ogre and don’t let them have their say, they won’t say anything.  That would be truly scary.  Teach them to talk back to you and you’ll hear about the process variables you didn’t think of and the ideas for alternative procedures that are tempting them to deviate, and you’ll find out the level of compliance with procedures that you’re really getting.

Talk it over and you may get agreement on a single way to do things.  If not, why would you expect to get the same results every time?

Layer 2 – “Too Many Chiefs Broil the Indian”

Maybe not my favorite way to make a stew, but when the recipe gets confused, it’s the workers that take the heat.

Maybe you’ve got too many Chiefs over-riding each other with “priorities”.  Maybe they’re not even talking to each other.  Maybe the equipment used varies with the size of the order.  Maybe the suppliers change depending on who offered the best price on raw materials this month.  Or maybe each mechanical problem is “fixed” by adjusting the process.

If the operators don’t get the same instructions each time, why expect the same results?

Layer 3 – “It Happens”

To misquote a famous fictitious moron marathoner who repeatedly found his place in history, “It happens.”

If you’re reading this, you probably have a title like “Quality Control Engineer”, “Project Manager”, “Troubleshooter”, “Duck Herder”, “Warden of the Fuzzy Messes That Have Been Swept Under the Rug”, or “Lord Master of Chaos”.  They all mean the same thing.  You probably already know that there are two kinds of machines, those that are currently running and those that have been bypassed.

The denizens of the Ivory Towers would have you take at least thirty data points, make a histogram to spot the sample that is outside the bell curve, hike up your shining armor, and go out to slay the dragon.

Have you ever seen a process work perfectly 29 times out of 30?  The swamp is FULL of dragons.  And if you get lab work done for the COA once per lot on a product you run quarterly, it will be seven and half years (and quite a few bungee bosses) before you have enough data to make the histogram.

A good mechanic will LISTEN to a car run, and tell you what’s wrong with it.  Your operators will develop a FEEL for a process and can say that something’s not right.  Do you listen to THEM?  This is your first red flag that something needs fixed.  What is the mechanism for capturing their observations?

Forget about specifications for the moment, how do you spot CHANGES?  You need to find a way to constantly monitor the most sensitive read-outs of your process status that you can.  Know what’s happening, and know the difference between what was happening when THIS pile was made, and when THAT pile was made.

Only then are you in a position to know if blending this odd bit in with the rest will save the odd bit or ruin the whole batch.  Life is full of choices.

Layer 4 – “What Do They Use This Stuff For, Anyway?”

It may not be your cup of tea, but one Ogre’s garbage is another one’s fillet mignon.  Face it, you were always going to sell this stuff from the get-go; you were never planning on using it yourself.

Hopefully, somebody out there is using your product for something, and hopefully is trying to make the results better.  Maybe they’re just learning how your product affects their results, or maybe they are doing something new with it.

Without knowing the application, you can’t possibly know what characteristics of your product matter – and of course the customer isn’t going to teach YOU how to run THEIR business.  They only have a fighting chance if your product is consistent.

Then, they ask you to make it better without changing anything.

Obviously impossible, but acting like an Ogre won’t help.  You’ve got to find out what matters to them and what doesn’t, even if you have to pound it out of them….  Then the impossible request becomes an exercise in pushing and tugging in different directions until you get it to fit.

But whatever you do, don’t fall into the trap of saying it’s the “new improved” version of the same thing, even if the advertisements must read that way.  Give it a new name; call a Troll a Troll and an Ogre an Ogre.  Don’t let them confuse you.

Layer 5 – “Ogre Not Lost, Swamp Moved.”

So you’ve got your foot in the door, and you want to get the rest of you in?  In other words, your product meets the customer’s needs some of the time, but just as often you’re stuck outside specs?  Why can’t you just make stuff that looks like that first foot, and not have to carry all the rest of that baggage.

If your customer is an Ogre, they might be tempted to chop off the parts of you hanging out the door and call it a deal, but you probably can’t afford to.  It’s a lot harder to make yourself smaller than to just push on through the door.

Nuts and bolts – moving the average is the quickest payoff.  First thing you do is a big multivariable study on how each of the outputs affects each of the end results, then calculate how you should…..  oops!

Not interested, you say?  Can’t afford the time or money?  Then you need to make the acquaintance of a little elf named Serendipity.  Serendipity will tell you that the fact is, you’ve ALREADY done all these experiments.  Every time something went wrong, a supplier changed, a setting drifted, you got a different result.  If you kept good records (You do don’t you?  You don’t burn the evidence to protect yourself, do you?) … you should be able to figure out which variables matter, and what happens when they change.  Every time your process went out of control and deviated, the goose just laid a golden egg.

When the customer says “we liked this lot better”, figure out what was different and give them even more of that, on average.  Brute force does wonders, as every Ogre knows.

Layer 6 – “You want me to squeeze into WHAT?”

Face it, we’re big.  We’re ugly.  We come with warts and pimples.  Sometimes our customers need better.  Just kissing the slimy green thing (and I’m not referring to the customer here, or management) doesn’t always turn it into a prince.

In the previous layer of peeling the onion, you may have gotten to where you want to be centered, but bits of you still stick out here and there.  You got in the door, but can’t get through the window – and all your body parts are attached.  In other words, the fluctuations aren’t “special cause events” because they’re built into the process.  Temperature and moisture are controlled only so well, given the change of seasons and time of day.  The machine settings drift or differ from machine to machine.  Power delivered depends on what else is running off the electrical grid.  Raw materials vary, and so on.

Now is the time to “pan for gold” – track down each of the factors in your system and follow it back to the source, to determine WHICH source of noise dominates in the output of concern.  Then, clamp down on that one.  This is long hard work, but it’s the only way to cut an Ogre down to size and make it fit through the customer’s window.

Layer 7 – “The Sun Never Sets In The Swamp”

Eternal Vigilance.  There are alligators out there, and the landscape never stays the same.

Enough said?  Get back to work!

You need to think about each product and figure out for yourself at which layer of the onion it is residing.  Then you’ve got a chance.  Good luck!

Dr. Gerald Keep