The Unbelievable Agony of Launching a True, Serial-manufactured Product


It amazes me that so many people in industry simply don't appreciate this problem.  I admit that I may be out of date by a decade or two, but I rather doubt that this issue has gone away.  You can be told all about this issue in case studies and lectures, but until you have been through it personally, you just don't understand it.  You have to have scars all over your back before the profoundness of this problem really sinks in.


My background is in discrete manufacturing rather than the process industries, so be careful not to over-generalize what I have to say.  But I suspect a lot of this does apply to process industries as well.  Also, my experience doesn't not include automotive-style production lines, although I believe the problem is, if anything, worse there.  The challenge is to take a product whose feasibility is known and design it, manufacture it in quantity and then maintain it in the field.  If you have already built one as a demonstration and know it works, how hard can that be?  Damn hard.


The challenge is that you now need to build and maintain a great many of these products without the involvement of the engineers who designed them.  If you are only building a few, when problems arise you can call the engineers to come fix them.  When the quantity gets too large this leads to exhaustion of the engineering team or simply overwhelms the engineering resources available.  Therefore, the serial manufacture of the product in quantity and its maintenance in the field must be done by personnel who did not do the design.


Conceptually, there are three different phases to be gone through:


1. Early in the development of the product (or even during an exploratory phase) a version of the product is built whose purpose is to prove the feasibility of the product beyond all doubt.  It works, it does more or less what the product will be expected to do, and it is sufficiently like the product to be able to answer all questions of feasibility.  In my experience, these product versions were called "engineering models".  While they work to some degree, they are not built to the product design (which likely doesn't exist yet at this early stage) and are not built by the production processes which will be used later on to produce the product in quantity.  If you only need one instance of the product, then this engineering model will be all there is.

2. At a later stage in the product development, another version will be built.  This one, when completed, will accurately represent the design of the intended product, which will, by then, be basically fully designed and documented.  This model will perform like the intended product and will be just like the intended product, except perhaps in some appearance aspects.  But the process by which it has been made is uncontrolled.  This model represents the product in detail, but does not represent the process by which the product will be built in quantity.  Indeed, if the manufacturing process is not important --perhaps you only will be building a few examples of the product -- this model is actually the first example of the finished product.  Often such a model is called a "prototype".


3. A third type of instantiation of the product is built to the design prints and is built by the manufacturing processes that are intended for full-scale production.  It proves that the product works and that the manufacturing processes, procedures, training and documentation work well enough to product a working product.  The first models born this way are often referred to as "pilots".  They are, in fact, the first production examples and show that continued production can be done when the business requires it.


I expect that different fields of work use different terms for these several stages, but the concepts have to be the same.  In fact, it seems likely that some industries will have several stages of proving out the production processes, and thus will have several stages equivalent to pilot models.


The reason that all this is worth going through is that, in my experience, there is often confusion and misunderstanding about what an early model of a product actually represents.  Just because you have built one (or a few) of something, and even maintained it in the field, does not mean that you know how to produce and maintain it in quantity.  It isn't just more of the same; it is a wholly different requirement and will affect both the design and manufacturing profoundly.


And, having said all the above, even if you have gotten the concepts straight and have carried them out with your best talent and effort, you will then find that there are an incredible multitude of details, arising one after the other, that will need to be corrected in order to get to a satisfactory production status.  As I said at the start, if you haven't been through it, you just cannot imagine or appreciate the effort and patience which will be involved.