Book review of the month

Since this is my first one of these, the title of this post is rather optimistic. However, I am trying to trick myself into reading more, so ssshhhh...

The honor as the inaugural recipient of my book of the month review prize is Stephen Toulmin's "The Uses of Argument".

I recently shared some thoughts with a colleague who happens to also be a graduate student. I cannot think of a better book recommendation for any graduate student and for any researcher than Toulmin's excellent book.

I wish I would have read this book, or any of Toulmin's books, many years earlier. I was introduced to this book by Dr Yair Levi when I took a doctoral course in systems engineering (out of all things). From the researcher/student perspective, it was life-changing. Toulmin's book is quite philosophical, and it certainly reminded me what the "Ph" in PhD represents.

The book presents a structural model for presenting arguments based on facts, backing, and reason, among other concepts. The book is full of colorful examples and it is very readable. Do not expect to follow it like a recipe book. Simply absorb and embrace the philosophy and the environment presented by the book and try to consciously apply those concepts to how you formulate arguments.

In my opinion, one of the most important steps when embarking on any research, or enterprise for that matter, is properly stating the problem. Therefore, understanding how to logically model an argument is key to driving things forward.

After reading this book I found myself applying these concepts to my own everyday job. I will let you in on some of my secret methods. I actually developed my own simple technique based on Toulmin's model for formulating arguments by using yellow stickies. My wife likes to decorate my office walls with maps, etc. and those just happen to be the perfect surfaces to brainstorm and "frame" (get it?) some ideas:

The only minor issue I have with this book is that it reads like a collection of lecture notes. That is, there is no real beginning, middle, and end.  Well, there is a progression to the ideas presented but, like in a textbook, the heavy lifting happens in the reader's head. This is a very minor point, because the book is a must-read for any researcher or professional. I highly recommend it.

Bees Do It

I am sure Ella Fitzgerald did not think she meant additive manufacturing, but that is a big part of what bees do.

Honeycomb map of the United States, created with the help of bees (credit Ri Ren, www.borepanda.com)

Additive manufacturing is the process of building something by adding material. Normally, manufacturing is a subtractive process, that is that material is removed, such as in milling, machining, boring, etc. Additive manufacturing is the opposite, where a product is created layer by layer. The popular name for this process these days is "3D printing".

There is a lot of interest in the subject of 3D printing these days, but the body of research is not very deep. I have been observing the hype from a skeptical distance for a few years. 3D printing has been touted as the cure for all kinds of maladies in every industry and process.

One of my issues is that many practitioners fail to see 3D printing for what it really is, a manufacturing process.

I once had to talk a DoD logistician away from the idea of using 3D printing to replace inventory items on the shelf. That is not necessarily a bad idea. However, inventory is near the end of the chain and, once we consider that we would be squeezing all the logistics elements into that one near-end point in the supply chain, it becomes obvious that the idea is more complicated than it seems. For example, quality control, training, technical documentation, who is the technical authority that certifies or qualifies the product to be introduced to the weapon system? On and on...

I would like to someday conduct finish a proper research study on the impact of 3D printing in logistics, and I have started some work on that already. My logistics instincts tell me that introducing 3D printing in the manufacturing or sourcing node of the supply chain makes the most sense initially, but we'll see where the research takes us.

Preferred method of measuring preference (and how to measure it, preferrably)

I was recently presented with an interesting problem. If a policy states that a particular method is preferred over all others, what would a metric to measure compliance look like?

I have been teaching Decision Support Systems for 6 years now and, academically, this is what my students call these days a "no-brainer". If preference implies that we want something maximized (e.g. money), then 51 dollars is preferred to 49 dollars, because 51 > 49. If the preference implies minimization (e.g. our tax rate), then 49% is preferred to 51%, because 49 < 51 (or, more accurately -49 > -51). This would be like the "popular vote" in a two-candidate election expressing the preference of the voters.

So I put my explanation in writing, with what I thought was a useful link to Wikipedia, rather than an overly academic reference to "Rational Choice" by Itzhak Gilboa (for instance). Big mistake. Turns out that practitioners are very suspicious of anything from Wikipedia and, by association, anything that agrees with it. So, what I thought should have been a slam dunk turned into an all-out  squirrel chase.

That gave me time to pause. Was my explanation wrong?

Sometimes, in math, in trying to explain something that we think is very obvious, we complicate things more.

As it turns out, it depends. If we consider the practitioners' side, this whole model depends on how it is implemented.

Philosophically, the model depends on knowing what constitutes preference and what the domain of choices is (the consumption set, in decision science-speak).  So, to fully develop the metric, we must define:

• How is preference demonstrated? Is it by the outcome of the decision or by including the preferred outcome in the decision process? For example, we prefer to win the lottery every week, but do we demonstrate that by winning most of the time or by buying tickets consistently?
• What are the choices? Is it a binary choice (A or B)? or is it a choice among a larger set? What does "over all others" mean? Are we lumping all the non-preferred outcomes as one single choice?
• Is this a subjective measure? If so, how is subjectivity operationalized in a way that can be detected when conducting business analysis? For example, it could be as simple as a person signing a document saying that they prefer unicorns to puppies as pets, even though the preferred choice is totally unachievable (or is it?).

As you can see, nothing is ever easy.

Material Management Life Cycle

One of the risks that we all must face as logisticians is in understanding and acknowledging the mission that our processes are designed to support within our respective organizations.

Many years ago, when I was in graduate school, I was exposed to the concept of Michael Porter's Value Chain, where the primary activities of an organization or business are identified as activities that create value, while secondary activities are identified as those that support the organization.

Michael Porter's Value Chain Model

Lately, I have been exposed to various different perspectives on what it means to manage material in DoD. As the saying goes, it matters where you sit. I have heard arguments for a strict compliance model with no value-creating business application (unique item management). I have also heard arguments for changing business models in order to make compliance easier. When put into the perspective of the old functional silos way of thinking, I suppose each of those arguments makes sense in their own self-contained world.

However, those silos of mentality are what we are trying to get away from. Ironically, one philosophy that could help is a return to fundamentals. Some modern business people have adopted a "first principles" concept from philosophy and physics, and that is exactly the kind of old-new thinking that is needed in order to effect the positive transformation that we seek in DoD. 

Material Management Life Cycle

To apply the Michael Porter value chain model, we can argue that the fundamental reason to manage material is to create a product or deliver a service.  I put together the above graph, adapted from various other models, to show how material moves in DoD, in most cases. The graph is divided into four quadrants or sections. The darker the arrow, in my estimation, the more critical that material management is to current or future operations.  The darkest section (let's call it quadrant 1) is where the bulk of the material management effort takes place.

So what does that mean? For the time being, just a point of discussion. The supporting activities in the DoD value chain that relate to material management need to consider where in the value chain their efforts are most effective. There is value in accounting, procurement, and information technology, but not at the cost of negatively impacting the missions that they support.

Decision Theory Under Uncertainty (Rational Choice)

I guess enough time has passed that I can confess that my previous post titled "Mandates and Pragmatism" was an allegory about the IUID mandate, and other mandates like it (if the shoe fits..).

When one steps back and asks enough "why" questions, it becomes evident that the problem that we are trying to solve is the problem of a mandate.  It is the "because I said so" rationale of our parents. Therefore, removing the mandate solves the problem. No? So, go ahead, kids, and eat your dessert first. Tell your parents I said so.

During one particularly futile discussion about the subject at an IUID conference I took to doodling in my notebook and produced the following mathematical model.

You see, I spent many frustrating months studying Izhak Gilboa's book "Decision Theory Under Uncertainty", so I figured that I would finally put some of that knowledge to good use. Gilboa is one of the living legends in the field of decision theory. Another one of his famous works is "Rational Choice".  I must warn you that his works are no light reading. One of the things I usually found when reading Gilboa was that it would take me a lot of effort to get through his mathematical models and proofs, only to arrive at a "duh" moment. So you mean that green marbles are not blue?

To summarize my doodle, the model is actually a mathematical representation using decision theory notation of a very fundamental issue: what is the rational choice?  If the set of choices are Navy ERP, IUID alone, or a combination of the two?

The value of the utility function was in terms of FIAR compliance.

The conclusion of my doodling is that concentrating on ensuring that the Accountable Property System of Record (APSR) is FIAR compliant would be a more rational choice than trying to comply with both the IUID mandate and APSR FIAR compliance, since IUID alone would never provide 100% FIAR compliance. Like my kids would say... uhmmm.... duh! Once we put it like that... who needs math?

What's the formula for statistical sampling?

Statistical sampling, for some reason, is one of those concepts that gives some people a lot of trouble. There is a book by Daniel Kahneman titled "Thinking, Fast and Slow" that I always recommend. One of the concepts Kahneman likes to write about is System 1 thinking vs. System 2 thinking. I suppose that those who feel comfortable when dealing with uncertainty and probabilistic models are people who are good system 2 thinkers.

So, what is statistical sampling? To keep things simple, it may be better that we start with an example.

Suppose that we work at a light bulb factory that makes millions of light bulbs per day and are in charge of testing the light bulbs. We could test for many things but, to keep things simple, let's say that we are testing whether a light bulb can survive a drop from 3 feet onto concrete. I am not sure why anyone would want that, but it sounds like fun.

There is one sure way to ensure that we can test this and be absolutely certain that we know what percentage of our light bulbs could pass the test. That is, by dropping every single light bulb on a concrete floor.

The problem with that approach is that we need the light bulbs for other things, such as selling them to make a profit.

Statistics gives us a way to test the light bulbs and have less broken glass to sweep.

Using statistical sampling, we could take a small sample of light bulbs and use them to represent the entire batch. We just need to follow some simple rules:

• We must choose the light bulbs in our sample at random
• Each light bulb must have an equal chance of being selected
• We must select a large enough sample size... (more on that in another post)

So, if we sampled one thousand light bulbs at random and 100 of them broke, statistics allows us to say that 90% of all light bulbs we produce can survive a fall of 3 feet onto a concrete floor.

What scares people off is what follows. There is a probabilistic nature to statistical processes. So the result of our drop test is not really 90% - it is actually "around 90%", in other words a range of values determined by something called a "confidence interval".

We are not going to get that deep into statistical sampling in this post, but I wanted to start things off with a very simple scenario that shows the benefit of statistical sampling.

The Second Law of Thermodynamics and Logistics

In previous articles, I have covered the basics of material identification, cataloging systems, and even a slight tangent into counting methods. So, naturally, we are ready to talk physics...

Well, not really physics, but I do want to introduce the concept of entropy in a business scenario. Let's start with a definition from Wikipedia:

"The second law of thermodynamics states that the total entropy of an isolated system can only increase over time. It can remain constant in ideal cases where the system is in a steady state (equilibrium) or undergoing a reversible process."

What that really means is that the toothpaste in the tube is in an ordered state but, as soon as you change the system (squeeze it, or just throw a 2 year-old in the room), the toothpaste becomes in a disordered state (or a big mess). Systems tend to transition from order to disorder, which is why a pile of rocks does not spontaneously become a wall but, leave a wall standing long enough and it will become a pile of rocks.

Ok, so what the heck does this have to do with logistics? Actually, a lot.

Let's say that we have a warehouse where our inventory accuracy is a perfect 100%. Should we even bother to conduct physical inventories ever again? After all, we are perfect! Not so fast. Let's call that scenario our perfect tube of toothpaste, or a beautiful wall.

Now for a painful segway - back when I was working on my first PhD dissertation (yes, first), my topic of research was the impact of information quality on the supply chain. After waiting way too long to say something, my advisor felt that was not an important enough topic and made me start a new dissertation after two years of work, but I digress. I will not bore you with academic background, but you can research the works of Fangruo Chen, Simchi-Levi, and others who have written extensively on the subject. Even Claude Shannon wrote about information entropy. In short, a percentage of all business transactions result in errors - sometimes as high as 2% of all transactions result in some error.

Welcome to the second law of thermodynamics in logistics.

Someone might say "so it is just 2%, that is still a 98% accuracy".  Not really. Keep in mind that every item in inventory had to undergo a number of transactions to get into our inventory records. Catalog record master data had to be created, procurement, material identification, receipt, put-away, etc.

Now you can see our dilemma. If we don't do anything, the system will continue to turn our beautiful wall into a pile of rocks. The only way to fight this business entropy is, to say it in physics terms, to spend energy. We must conduct physical inventories and we must reconcile the "physical record" with the information record, and we must do this constantly, forever.

The NSN (Part 2)

In this post, we will expand on the NSN concept that we introduced in part 1.

Let us start with a simple mental exercise. Can you tell what type of number this is?

(866)555-5555

If you said a telephone number, you are correct. We have also been conditioned to recognize even more obscure numbers such as 012-01-01234 (social security number), or 12345-1234 (zip +4). If someone asked us to pick out the area code or prefix from the telephone number above, most of us would know the answer.

The NSN, for those with even a vague familiarity with DoD logistics is also just as easily recognizable.

Here is an example: 5320-01-4355075.  That is the 13-digit NSN for a rivet.

Like a telephone number, the NSN is a combination of meaningful sections (data elements).

Let's pause, for a moment, and bring up some history. Before NSNs, there were FSNs (Federal Stock Numbers). At one point, everything in the DoD supply chain was identifiable with a 7-digit Federal Item Identification Number (FIIN). FSNs were officially recognized in DoD from 1955 to 1974. So, do not be surprised to still find some legacy language using the term "Federal Stock Number". The bottom line is, there is no such thing as FSNs these days and, no, it is not an interchangeable term for National Stock Number.

However, the NSN does build on the original data elements of the FSN.

Let's break down the NSN in the example above into its components:

"5320" - Federal Supply Classification (FSC). This is the descriptive category of this item. One way to think about it is, if we were in a department store, which section would we go to find this item.  To use the telephone analogy, you can also think of this as the area code.

     Note: the FSC can be further broken down into "Group" and "Class". In this example, "53" is the group and "20" is the class. However, this level of detail has gone into disuse now that we rarely see printed catalogs, since the FSC is only useful for cataloging or grouping items of the same product category. There are some exceptions for keying only on the group, such as the use of group 89 for food items, since those are of interest only to personnel working with food.

"01" - National Codification Bureau (NCB) code. This designates the country that assigned this NSN. Hint: 00 and 01 are the US.

"4355075" - This is no longer used on its own. This is the old FIIN. Instead, because other NATO countries might assign the same last 7 digits, today we use...
"01-435-5075" - this is the National Item Identification Number (NIIN), which consists of the NCB code and "the last 7" (nobody says FIIN anymore, although FLIS still refers to this as the IIN internally). Note that it is common practice to break up the last 7 digits with a dash, but that is only for readability and it has no significance whatsoever. The NIIN is what uniquely identifies an item in DoD catalogs.

So, to recap, the NIIN is the thing. That has a nice ring to it, doesn't it?