Quality & Material transformation

For nearly ten years, we note the systematic technical difficulties of industrialists in the transformation of materials within organizations of all sizes, in sectors yet as diverse such as cosmetics, food, pharmaceuticals, composites, coatings or waste treatment. The factors influencing conformity and variability often raise questions and the development of new products is often an obstacle course.

However, there is no shortage of methods for Quality and most of the players concerned by these difficulties have Quality systems and procedures in force. How can we explain then that the development or industrialization of new products can be systematically so costly in time and resources?, with a significant impact on time to market?

Would there be different levels of implementation of the Quality approach? Do certain practices fall outside the scope of continuous improvement? Are the technical aspects less obvious to tackle than the manuals suggest?

Over the years, it has gradually become necessary for us to look at the heart of quality methods and their presuppositions for the possible causes of their own limitations.

The universality of Quality

Originally, Quality was centered on Control in production. Then it gradually widened its field of action to become an Approach placing the satisfaction of customer needs at the center and widening the involvement of the different layers of organizations until it became a theme of Management, encompassing all dimensions of the activity of an organization [1]. Some methods have even taken the name of Total Quality.

Each method, on the strength of its successes - and probably concerned with widening its field ofapplication-, is often presented as universal, that is to say applicable to any industrial context. It is very difficult to find any explicit mention of possible limitations - on the part of their promoters at least... On the contrary, the failures or practical limits of implementation are often attributed to a poor understanding of the principles of the method or in their implementation [2].

However, it does not seem unreasonable to consider that any analysis, theory or method applies within a scope of validity that it is useful to seek to define in order to allow optimal implementation.

Process stability

Historically, quality methods are essentially derived from statistics, whose methods William Shewhart in particular introduced and popularized at the industrial level as part of his activity within Bell Laboratories in the 1920s [3] [4].

Shewhart starts from the observation that a manufacturing process is inevitably subject to more or less significant and controllable variability. When the variability leaves the attributes of interest of the product within the fixed tolerance intervals, the process is considered under control. Conversely, when these variabilities are outside the tolerance limits, the process is unstable ; depending on whether these variabilities are chronic or sporadic, the diagnosis is more or less severe on the instability of the process. Shewhart associates the notion of cause with the notion of variability. He defines the hazardous cause as the one whose impact leaves the process in its area of ​​control and the assignable cause as the one that needs to be identified and eliminated [4].

In the context of quality in production, the objective is thus to preserve, and if possible to improve, the stability of the process. The problem appears when the process becomes unstable, by an excessive deviation of the quality compared to the interval of tolerance. The entire history of Quality seems to be based on this principle.

The instability of development

There is, however, a context in which instability is not only inevitable, but highly necessary: ​​the development and industrialization of a new product. Indeed, when the characteristics of the targeted product are sufficiently different from the products usually produced by an organization, the process is by definition unstable since it is a matter of adjusting it - in some cases even developing it, or even invent- to give the product the required characteristics.

All sectors of the formulation or transformation of materials are concerned : food, cosmetics, pharmaceuticals, ink, coating, paint, ceramics, composites, petroleum, lubricants, etc.

This step is inherently problematic, costly in terms of time and resources, all the more complex in organizations with high turnover, with a serious impact on time to market. The phenomenon is no more a difficulty for SMEs than for large accounts.

From common factors to common causes

The observation of the similarities of difficulties between organizations of yet varied sectors is not without implication. In particular, he suggests that the common causes of difficulties could be found at the level of factors common to the actors concerned. Let's name a few.

The first of the common factors is related to the nature of the products and activities concerned : operations related to the treatment and transformation of matter, often neither completely liquid nor completely solid to powders, pasty materials, gel, emulsion, phase change products, etc.

A second common factor is that these actors develop new products or processes. This is not a standardized mono-product manufacturing activity, but on the contrary issues associated with the different stages from R&D to marketing.

A third common factor is that most actors proceed through these steps by the most traditional route possible : empiricism, that is to say by trial and error with the essential support of the know-how available in the teams. It is therefore a question of “doing”, as the cook tests his recipe and its variants until he arrives at a dish to his liking.

Thus, in these unstable phases, it is not a question of thinking about the resolution of a specific problem, but rather the systemic causes of a bundle of systematic problems.

Knowing that the usual Quality procedures have for the most part been developed in the context of assembly industries (telephony for Shewhart or Motorola, automotive for Toyota, weapons for Crosby, etc.), it seems useful to me to inspect whether certain traces of the specificities of assembly activities would not be buried at the heart of the methods, thus making their transposition to the technical issues of processing potentially problematic.

We planned it this way

Joseph Juran, one of the big names in Quality, puts us on the right track in the preface to his famous Handbook of Quality: “Numerous specific quality crises and problems have been traced the way in which quality was planned in the first place. In a sense, we planned it this way (Many specific quality issues and crises stem from the way quality was set up initially. To some extent, we organized it that way) [5].

"Universal sequence of Quality Planning" according to Juran, from [5]

If we look at the approach proposed by Juran, we see that he defines a "universal sequence of quality planning", illustrated in the figure above. In this approach, on the technical level, it is a question of designing the product then the process which will make it possible to manufacture it effectively. To achieve a process under control, the logic of stability requires setting the conditions linked on the one hand to the material, on the other hand to the process.

Sequence vs Reality

Yet if the logic of such a sequence seems implacable, its implementation for the technical issues of material transformation is far from being completely trivial. Indeed, the design and feasibility of the projected product – a new cream emulsion, a new dessert, a new composite, a new fat, etc. – requires a first concrete realization. This is the mission of R&D.

It involves transforming ingredients through processes to make a product with specific characteristics. In these transformations, material and process are as decisive as each other: cchange the materials, the product changes; change your process, or even the order of incorporation, the product is likely to change too. The culinary analogy is always enlightening on these questions. Imagine preparing a vinaigrette and think about the impact on the texture of the emulsion of the speed at which you stir a spoon or fork.

In the following stage, when a satisfactory version of the product is obtained on a laboratory scale -that is to say with certain means of preparation-, recourse to other means for manufacture on a larger scale does not does not go without saying. Clearly, it is not a question of assembling independent parts; product and process are strongly coupled. This is a feature of these so-called complex systems, which also makes them rich in application.

Most often, products are composed of several ingredients (often between 5 and 40), for which it is absolutely impossible from the independent properties of the ingredients to deduce those of the finished product. Thus, in the absence of computer-aided design tools used in the vast majority of mechanical and assembly industries, the use of empirical approaches seems to be the only approach -we will however see when we present our own methodology that more effective alternative approaches can be developed.

Measurement: essential but invisible

An additional element of analysis concerns the measure, which Juran indicates in the commentary of the universal sequence that it concerns all the stages. But what measure is it? In his work, and most authors do the same, Juran mainly refers to measurement by sensor, that is to say the measurement of “simple” properties.

In material transformation, simple properties are often insufficient. The properties of both inputs and finished products give no information on the intermediate states, which are crucial in the stages of material transformation..

On the contrary, the measurements as usually practiced in the field of material transformation even tend to consider identical ingredients or products whose differences are likely to cause notable variations in process as well as in application.

I had the opportunity to address these issues around the notions of viscosity in previous articles; the behavior of powders, for example, could be the subject of an equivalent demonstration. On these aspects, scientific knowledge has evolved considerably, much faster than measurement standards.

It should also be noted that in all approaches based on statistics, measurement is a presupposition, since it is from it that the data come. No relevant processing with erroneous or evasive measurements - in particular those which struggle to distinguish two products whose behaviors are in practice different. This point is crucial and yet very little questioned in the methodological manuals.

Measurement is absolutely essential, but it becomes invisible when it comes to processing data.

The misunderstanding of phenomena, a root cause of difficulty?

Thus, questions of development and industrialization of new products and processes in the material transformation sectors raise certain questions about the scope of application of the usual Quality approaches for technical questions.

In these stages of instability, no statistical reality can guide the action, so that the actors often find themselves forced to proceed by empiricism. In practice, it appears that the understanding of the phenomena at the origin of the observed variability is difficult to acquire.

On closer examination, most Quality methods consider such an understanding either as self-evident or as the result of statistical approaches in particular. It is quite surprising to note in most textbooks the complete absence of this notion of comprehension.

However, for several decades now, another great name in Quality, W. Edward Deming, has insisted on the need for "profound knowledge" which notably involves understanding variations and more generally the system of the organization [ 6]. For several years, the Food and Drug Administration has been promoting the Quality by Design method, which has the particularity of insisting on the need for prior knowledge.

Let us also quote a recent work by Quality by Design for the pharmaceutical industry, which in the introduction describes a situation which, in my view, goes far beyond the framework of the pharmaceutical industry: “Historically, an ultra-compliant approach had dominated the way the pharmaceutical industry operated, perhaps even threatening, wrongly, to potentially swamp the underlying science, rather than compliance being seen as a partnership with science.” [7] (Historically, an extreme approach to compliance has dominated pharmaceutical practices, possibly even at the expense of the underlying science, rather than establishing complementarity with the science.)

Such complementarity is certainly a major asset in a world where agility is becoming a major factor of competitiveness.

Projects

[1] C. Yang, The Evolution of Quality Concept and the related Quality Management, 2017, Intech

[2] G. Barouch S. Kleinhans, Learning from criticisms of quality management”, 2015,
International Journal of Quality and Service Sciences, Vol. 7 Iss 2/3 pp. 201 – 216,

[3] D. Bayart, The quantification of quality control in industry: a sociological and historical point of view. In: Rural economy. N°217, 1993. Quality in the food industry. p.p. 18-23

[4] WA Shewhart, Economic control quality of manufactured product, 1931, D. Van Nostrand Company

[5] J. Juran, Juran's Quality Handbook, 5th ed., 1999 (1st ed. 1951), McGraw-Hill Companies Inc.

[6] W. Edwards Deming, The New Economics for Industry, Government, Education, 1994, MIT Press

[7] W. Schlindwein & M. Gibson (editors), Pharmaceutical Quality by Design A Practical Approach, 2018, John Wiley & Sons Ltd