We Think that Alan Nichol hit the nail on the head with this article- It may be long, but logistics are INCREDIBLY complicated. This is one of the better explanations why many people who serve within the  logistical loop may only understand their small part in what has become a global expansive logistics machine.

If I were to take a poll of readers, collecting the answer to the question, “What one business process is obviously wasteful and also too difficult to fix,” I think I can predict what a common answer might be. Take a moment to consider your own answer to the question.


If you said, “My office processes,” that is a common problem place, or perhaps you know this author too well and you are trying to predict my train of thought. I’m looking for only a single process, or at least a single function where the challenge of eliminating waste seems too difficult. The collective office processes are too broad.

How about the shipping dock? I’m sure each of us has ordered something from a manufacturer, supplier, or on-line retailer and been surprised either by the seemingly exorbitant shipping fee, or by the oddly oversized, over-packaged box in which our items arrived. Have you ever gone a step further than shaking your head and shrugging your shoulders? Have you ever investigated the shipping fee?

I have, on a few occasions over the years, tried to figure out if I’m getting “ripped off” by a retailer or supplier by checking the true shipping cost of an item that I ordered. My few and unordered observations are various in their indications.

I have discovered that indeed a few businesses have apparently made a profit from the shipping fee, I have discovered more than one that apparently lose money on shipping, and a couple that somehow charged me exactly what it cost to ship a collection of items even though my on-line transaction took place before the items were pulled from a shelf and put in a box to ship. Consider that for a moment and you too might decide that such an accomplishment deserves some applause.

Does your own business struggle with the same problem of controlling expenses on the shipping dock? It’s not an easy problem to solve. I’ve tried. In fact, what sounds like it should be very simple has been, in my experience, one of the most complex conundrums to try to solve. So, because all of us have observed the problem either as consumers, customers, or process improvement people or process owners, let’s explore that problem to study how to solve the seemingly impossible process problem.

I have two very memorable experiences with trying to tackle the apparently out-of-control shipping costs problem, with two different companies. In one case, I was mentoring a newly trained Six Sigma Green Belt through his certification project where he chose to tackle the shipping dock costs. The second was a strange sequence of process improvement challenges that eventually led to hidden and uncontrolled costs at the shipping dock that a functional leader wanted solved.

Both experiences shared many of the same complexities and challenges, but also some incidents were unique and valuable lessons. For today’s story, I’ll blend both real experiences into a single hypothetical tale of problem solving for the sake of brevity and study. In other words, the events and challenges are real; the context will be changed to simplify the tale.

Start by going down to the shipping zone and observing what takes place. We can observe for several hours to get a feel for the variation and process challenges, or we can simply ask the process doers where the pain is hand have them show us. It gives us some perspective of what to look for and it saves time.

We observed a great many things, but a few things really stood out. When shipping a complete pallet of goods, there is a brilliantly engineered, and very specific, way that the boxes must be arranged to optimize how much materiel goes into a pallet of goods. Aside from some excess motion to shift boxes around in order to find the right ones to go in the right places, it is a very efficient process; not much opportunity there.

Then we are shown the real aggravation. It occurs in the small orders: replacement parts, specialty items, small direct orders, samples, or leftovers that didn’t fit in a complete pallet. The shipping department carries standard-sized boxes that facilitate the magical pallet puzzle. These boxes aren’t necessarily an appropriate size for the small and individual orders.

For the most part, the shipping boxes on-hand are designed to enclose 5 or 10 individually boxed items, so when we want to ship only one or two, the excess space in the shipping box must be filled with packing paper, peanuts, or air pillows. The extreme and embarrassing mismatch occurs when we witness the shipper placing a single screw pack (a small, sealed plastic bag with 4 screws for product installation) in a box big enough to hold 4 shoeboxes with a bunch of crumpled paper. He weighed the box, put a shipping label on it, and sent it off.

Yes, it cost more in shipping materials and shipping charges than the screw pack was worth. The shipper offered us an exasperated sigh and explained before we could even ask the question.

“We don’t have any smaller boxes.”

“Do you normally have smaller boxes? Are you just out of them right now?”


“Does this happen a lot, or did we just happen to see a strange event?”

“We do this almost every day.”

After long conversations and many questions we discovered the following challenges. There were no boxes for product smaller than the 5-to-a-box size. Smaller boxes were not kept in stock. They either cost the same price or more than the “standard” sizes they did keep, or they ran out or sat in inventory for too long. It wasn’t worth the trouble of keeping other sizes around.

We decided to dig through what little data was available to try and put a picture together of the variation that takes place in shipping. If we could predict the needs, we could order appropriate inventories of appropriately sized boxes and envelopes. As you might imagine, the data to do such an analysis was not available.

We asked the shippers to keep a log for us indicating what box was used for what orders. We came back a week later to collect our first week’s worth of data only to discover that we only had three lines of it. It turned out to be too big a chore and a disruption for the shippers to fill out our data sheet for every box. It took more time to fill out the sheet than to do the regular job.

After some, “Why didn’t you say so a week ago,” admonishments and some “Stop wasting my time; if it could be fixed I would have fixed it a long time ago,” dirty looks were exchanged, we finally came to a compromise. The shippers would double print shipping labels for us for the next month. With the shipping labels, we process improvement guys could reverse engineer what items were shipped for what cost, and in most cases extrapolate if the weight of the shipped item were significantly different from the weight of the item itself (over-packaging).

After collecting a couple of weeks of data we started to try and build a model of the problem. We constructed a histogram of the data and tried to fit a curve to it. Unfortunately, the data profile was extremely leptokurtic. That means that it had an extreme density of data points around the mean value and some relatively flat tails out to either side. In fact, the histogram distinctly resembled a rude hand gesture that communicated certain metaphorical significance as well.

I say that because, for the statistical layman, there is no good way to mathematically model a leptokurtic curve. We couldn’t predict the performance. So, we decided to break the curve into two pieces, then three, to see if we could model the individual portions of the curve, each side and the “finger” in the middle. Nope. The sides were distinctly platykurtic (shaped like a plateau with relatively low density of data points near the mean – the opposite of leptokurtic) to the point of being nearly uniform or perfectly random.

We hoped that another few weeks of data would change the profile’s structure, but it did not. It became clear that even with years of data, we would not be able to reliably predict the sides of the profile. The “finger” in the middle could be predicted based upon orders received, but the demand on either side of it was totally random. The department manager only had that well practiced, “I told you so,” look to offer us in the way of advice.

So, setting aside the vision of predicting box size inventory needs, we decided to try the problem from a different angle. Could we simply reduce the need for so much packing material? The individual products arrived to shipping in individual boxes, labeled with a stick-on label indicating the particular configuration of options and features inside. Could we not just redesign these boxes to be robust enough to ship without putting them inside a second shipping box?

The answer was yes, but the cost of the redesign and a more robust custom box of that size was more than the cost of the standard shipping box. In fact, the packaging engineers had already exhausted that train of thought in their effort to design the packaging and the boxes in the first place. It’s cheaper to use two boxes, than to design and purchase a ship-worthy individual box for each product. Seems unlikely, but it was the case.

There was another complication. To change the individual box we had to get the marketing function’s approval and they had very deliberately designed the appearance of the boxes. They weren’t on board with endorsing a cost reduction when we couldn’t even reliably predict what that cost reduction might be or clearly communicate how we thought we might achieve it.

At this point, my new Greenbelt looked frustrated enough to give up on Six Sigma forever, and was giving me that, “What now wise guy,” look, and I was getting worried about successfully mentoring him through his project. Still, we didn’t give up.

We briefly considered a universally adjustable packing solution to match up with a universally unpredictable shipping demand. We thought that raw cardboard could be cut according to a variety of templates to construct an appropriate size box on-demand, from raw materiel. It sounded good, but when we pretended to use the idea on real demands in shipping, we discovered we would create a great deal of scrap materiel, or that we would keep an excess of scrap material in inventory waiting for the small orders that could effectively use small pieces of raw cardboard.

We also had concerns about the ability to produce quality boxes from raw materiel without an enormous investment in capital equipment. It wasn’t practical. We had to keep looking for something else.

When we started digging into where the demands for shipped items were coming from, and how the materiel for shipping them was ordered, we opened up a real mess. For example, the screw packs generally weren’t ordered through customer service, they were dictated by warranty services.

A quick investigation revealed that an intermittent process problem had been allowing product to ship without the screw packs included. An interview with the process improvement person addressing it revealed a deeper truth. On larger installation jobs, the screw packs had a tendency to go missing, and rather than holding up customer jobs with due process, sales reps were ordering replacement screw packs through warranty channels. “Whatever keeps the customer happy.”

That was a whole mess we didn’t want to get involved in and we wished the process improvement guy already on the task, “good luck.” Potential tangent problems aside, we learned that there were multiple channels from which shipping demands would arrive, and that excepting the standard customer bulk order for distribution channel, none of them were predictable. Also, the various budgets didn’t account for the costs they incurred in shipping or weren’t held accountable, so the cost wasn’t connected to the decisions.

Connecting costs to channels of demand was something we could influence, and doing so might influence behavior. Finally, we were on the track of something we could control, but it wasn’t enough. At the same time we looked again at the shipping label data and the random variety of things to be shipped. We decided to look for the hinge points or break points in the cost profile for the shipped items.

We found we could successfully subgroup things into certain price brackets for shipping expense.  With a great deal of what-if replaying of the data we recently collected, we were able to devise a reasonable, limited set of shipping boxes and envelopes that could be kept on-hand without maintaining an excessive inventory of materiel. While some of the alternative size boxes didn’t cost more or less than the “standard” ones, managing weights and sizes did control some of the postage expense.

Part of the solution included pre-cut spacers that simulated the presence of an individual product box inside of standard shipping boxes that were more predictable in cost and less expensive (believe it or not) than the filler supplies we had been using, and did a better job of protecting the product. It also appeared to have some influence on customer complaints in our post-improvement follow up.

In the end we put a marginal dent in the cost of shipping items other than the bulk orders for distribution channels, which already demonstrated very little waste. What we learned is that what was apparently, obviously wasteful, was not as wasteful as it appeared, in the grand scheme of overall costs. Also, what waste was there was primarily driven by decisions that were disconnected from the costs.

In other words, those actions that caused us to ship screw packs (for example) did not suffer consequences in the form of a charge to the decision maker’s budget. While connecting consequences to the decisions did make people more sensitive to the costs, it didn’t necessarily change the decision and so providing a more cost-effective alternative for those necessities mitigated the impact.

When it was all said and done, process owners were pleased that we were able to improve their processes somewhat and remove some of their pain, and change agents were successful in making a measurable and significant impact to costs. We were more successful than process users expected, while less successful than those outside the process anticipated.

Naturally, the complexities of the problem were more convoluted that I took words to describe in this post while what I did share rivals the intrigue and interconnectedness of a good mystery thriller. So, what is the point of this long and convoluted tale? We can sum up the lessons learned at least as follows.

▪Not everything that appears wasteful is as wasteful as it appears

▪Data collection and analysis can be more painful than the process problem

▪An elegant solution isn’t always pragmatic

▪Fancy tools and statistics don’t always solve a problem, sometimes the best solutions are born out of common sense and “grunt work”

▪Even given the above three observations, a difficult and complex problem can still be significantly impacted if we persevere a keep looking at all the angles and for those elements we can influence or control. The solution may not be what we anticipated, but a practical improvement can be found.

The last bullet is obviously the one I want to emphasize most with this post. Don’t accept that a process problem cannot be solved. If it causes pain, it is a problem. Any problem can be mitigated or improved upon if we take the time and effort to really understand it and address what we can control. Sometimes, we just need to change the rules so that something we don’t control becomes something we can.

Some process problems seem like they should be simple, but turn out to be extraordinarily complex. When this happens, we have a tendency to declare it too complex or difficult to fix because our first instincts, tactics, or tools don’t find an acceptable answer. Don’t give up so easily. Dig deeper, break it down into smaller pieces, turn it over and look at it from different angles until you do find a way to improve performance. Challenge the notion that a problem is too complex to solve.

Stay wise, friends.

via Too Complex To Solve.