Sides of a Coin

Most controversies would soon be ended, if those engaged in them would first accurately define their terms, and then adhere to their definitions. – Tryon Edwards

simple: easy to understand, deal with, use, etc.

simplistic: characterized by extreme simplism; oversimplified

complex: composed of many interconnected parts; compound; composite

complicated: difficult to analyze, understand, explain, etc.

Such small distinctions.

Such calamity when those distinctions are not properly acknowledged, particularly when applied to the supply chain.

Let’s start here:

  • Problems that are complex can usually broken down into smaller problems, each of which can be simple
  • Problems that are complicated are often intractable – they can’t easily be broken into chunks and require some sort of breakthrough to be solved (think peace in the Middle East)
  • Problems that are simple – well, they’re not really problems at all, are they?

Now here’s where – in my experience at least – retailers start running into trouble.

Problem #1: Seeing complexity as complication

If you’ve ever read articles about the challenges of planning the supply chain, you’ve probably seen a diagram or two that looks something like this (if not worse):

The implication is clear: The is a vast, sophisticated web of relationships between and among various entities. Not only that, but the problem becomes more mind boggling the closer you get to the customer, as depicted below:

I mean, look at all those arrows!

To be sure, the planning problem is large in terms of the number of items, locations and source/destination relationships. The combinations can get into the tens – if not hundreds – of millions.

But is it truly an enormous and complicated problem or merely an enormous number of simple problems?

In contradiction to the popular proverb, sometimes it’s necessary to focus on the trees, not the forest.

How you consider the problem will in large part determine your level of success at solving it. Complication is often a self inflicted wound (more on that later).

Problem #2: Seeing simplistic as simple

Just as overcomplication leads to trouble, so does oversimplification.

Best summed up by Albert Einstein:

People love this quote because they feel it very neatly sums up Occam’s Razor. Then they further reduce it to a design philosophy like “the simpler, the better”.

But the real power in that quote doesn’t reside in the first part (“Everything should be made as simple as possible.”), rather in the second: But not simpler.

In a planning construct, I’ve most often seen this line crossed when the topic is for slow selling item/stores. Demand forecasting as a discipline originated in manufacturing where volumes are high and SKU counts are low. As a consequence, most mature forecasting methods are designed to predict time series with continuous demand.

When you look at consumer demand at item/store level, a significant portion of those demand streams (often more than 70% for hardlines and specialty retailers) sell fewer than one unit per week.

So we have intermittent demand streams and tools that only really work well with continuous demand streams. So what’s the solution?

One of the most common responses to that question is: “Screw it. It’s not even worth those items. Just set a minimum and maximum stock value for ordering and move on.”

That position is certainly not difficult to understand, but is this really keeping it simple or being overly simplistic?

Here are some follow-up questions:

  • What if a particular item is fast selling at some locations and slow selling at others? Would you forecast it at some locations but not at others? How would a demand planner manage that?
  • If not every item in every location is being forecasted, then how do plan future replenishment volumes? Not just at the stores, but at DCs as well?
  • If not every item is forecasted, then it’s not possible to do rollups for reporting and analysis, so how do you fill in the gaps?

Because a simplistic solution was proposed that only addresses part of a complex problem, we have created a problem that is complicated.

This is akin to looking at a single tree without even recognizing that there are other trees nearby, let alone seeing a forest.

Things are only as complicated as you choose to make them, but they also may not be as simple as they seem.

Leave a Reply

Your email address will not be published.