In one of the most ridiculous manufacturing schemes I have ever read about, Intel explains their "Copy Exactly" strategy of optimizing manufacturing. An Intel Plant Manager describes Copy Exactly: "It’s not just [that] there’s a specification or a recipe or a program you put into a machine," he says. "It also is what the human being does and how they interact with the machine."
This is the McDonald’s of manufacturing. It is making McChips – straight from Ray Kroc’s instruction manual. All that remains is for Intel to toss out their "Leap Ahead" slogan and adopt Quality, Service, Value and Friendliness. And they ought to scrap the "Our success depends on talented employees around the world who are passionate about technology. Are you ready to make the most of your mind?" nonsense on their web site while they are overhauling things.
Kroc (that’s him smiling to the left), of course, created Hamburger University, where every detail involved in slapping a piece of meat between two pieces of bread is honed to an art form before it is rolled out for teenagers around the world to duplicate.
Andy Grove and his #2 man, Craig Barrett, must have done a lot of power lunching at McDonald’s to have cooked up Copy Exactly. Only they are not trying to get teenagers under control. Their philosophy is applied to adults.
"Intel Corp., the world’s biggest chip maker, is unique in the way it rolls out new manufacturing methods, perfecting it in a laboratory and then painstakingly duplicating it at factories around the world."
"Under Copy Exactly, researchers spend more than four years perfecting a new manufacturing technique in one of Intel’s development factories in Hillsboro, Ore. Once they are satisfied with the results, they work to meticulously import every last detail to half a dozen or so chip factories around the world."
This scheme is the polar opposite of engaging and involving employees. It stands continuous improvement on its ear. Engineers working in remote laboratories decide everything down to what color gloves the workers will wear to the paint scheme on the walls. Ostensibly it is to be sure they control every variable.
So how’s it working? Andy Grove and Intel neither know nor do they care. 
(The thoughtful gent to the right is Mr. Grove, the man in charge of the whole scheme.) Richard Doherty, research director at the Envisioneering Group, says the technique is "rooted in Intel’s corporate culture and there’s no scientific evidence that it gives the company an edge". Translated, this means, "We don’t know if it does any good and we don’t care how degrading it is to our employees to be told to shut off their brains and sleepwalk through their jobs like Zombies, it is just the kind of guys we are.’ Some culture you guys got cookin’ there at Intel.
Intel proudly boasts that they have achieved "pretty good" status as manufacturers. I would think that the journey from ‘pretty good’ to even ‘very good’ will be like climbing Mount Everest in high heels so long as they have disengaged the minds of the entire production workforce. Going all the way to ‘excellent’ or ‘world class’ is out of the question.
"All Intel cares about is that a chip comes out and is electronically the same as every other chip coming out." – Intel philosophy
"It’s the job of Hamburger University to ensure that those results get replicated each day in every country where McDonald’s operates" – McDonald’s philosophy
Can I get fries with those chips, Andy?
Get Some Data says
Where is your data to show “all employees” are disengaged? Have you ever visited an Intel site or talked to Intel employees?
Bill Waddell says
Mr. Get Some Data,
I suggest you get some grammar. Quotation marks go around direct quotes. Nowhere did I write that all employees are disengaged. In fact, I wrote, “they have disengaged the minds of the entire production workforce”. Quite clearly the engineers back in the lab are deeply engaged. The recipients of their detailed job instructions are the ones who are disengaged.
And, yes, I have visited Intel sites and know some Intel employees.
Micheal Gardner says
It would seem the ultimate question would be whether this system works well for Intel or not. I don’t know for sure, but I suspect it does not. My understanding of their method is that all processors are made to be the top of the line–what are they up to now, Pentium V? However, when they go through final testing, many do not pass the spec requirements and become “lesser” processors, like the Celeron. If everything in the manufacturing process is exactly copied from the best practice, this would not happen. It would seem, then, that the “copy exactly” method is not actually doing what Intel wants it to do.
I’ve never visited an Intel plant, but I know other circuit and chip manufacturing operations are mostly automated. Robots are pretty good at doing everything the exact same way every time. I have to think Intel’s plants are highly automated as well, so what is the point of the “copy exactly” system?
Bill Waddell says
You’re right, Mike, it does not seem to be working for them too well. My take on it is that, because their failure to control some variables in the production process in the past caused problems, they launched a gross over-reaction in order to try to control every variable.
It is a scheme Frederick W. Taylor would have loved – total control of every move production operators make from a remote location filled with smart people.
The problem, obviously, is that any operator generated ideas for improvement become extremely difficult, if not impossible to implement. What they may have gained in quality control they more than gave up in loss of employee contribution to continuous improvement of cost and quality.
It seems to me that they fell into a familiar trap. They bet the ranch on a relative handful of people in the lab being smarter than the thousands of people in the production workforce. While their lab rats may well be smarter than any production worker individually, the lab rats are mental midgets compared to the collective knowledge and creativity of the production workforce.
I imagine the Intel execs think they have this covered with some sort of convoluted system in which a production worker in India can make a suggestion and have it go up, across and through all sorts of channels until it reaches the lab rats, who then evaluate it and decide yea or nay. If so, that ‘suggestion system’ has been in place at General Motors since Alfred Sloan took over in 1922 and has never worked.
Ralf Reinhardt says
I think one must consider the technology used. Intels way of exactly copying their fabs is not unusual in this kind of manufacturing process. The reason is that creating chips with sub nanometer precision by incurring impurities of 10^-10 to extremly pure silicon is a process that can be disturbed by very small effects coming from very remote sources. In comparison to automobile production the search for causes is rather difficult and very costly.
Just copying an existing fab is dissatisfying for any QM educated technician but the circumstances make standard QM techniques a costly endevour.