Taking my own advice

I recently wrote a program that calculates lead time, filtering out unreliable data statistically. My objective is to generate a reliable lead time, not a perfect predictor but sound enough on which to base forecasting and replenishment. Since it’s not perfect, people who use the results will need to understand how it works. But – and this is the catch – they don’t really want to know how (that is, they don’t want to have to think about standard deviations), while they do need to be able to interpret credible results (which are sometimes not intuitive), and distinguish them from non-credible results.

Just a few days ago, I wrote here that the challenge is to embed process change within the social context we inhabit. How do I create a narrative or even entertain (!!), while implementing this tool? Just thinking about it takes me down a completely different path that my old-school teaching mind imagines (drawing equations with Greek symbols on a whiteboard…what was I thinking?).

Now I’m thinking about it from the learner’s perspective: why would I need this information? Why would I want this information? How will my life be better if I know this? How could I explain this to someone else in my department? The task now is to discover the human story in the solution, and tell it memorably.

This is much harder than writing the code.

Stock-outs

I started looking into stock-out occurrences yesterday. The high incident rate on A items alone just stunned me – wow, what an opportunity! If a SKU performs well enough to drive 80% of our business in a category, despite the fact it’s not always in stock, we don’t even know its potential.

I used to work in the fastener industry, where you knew the potential of your top sellers: there’s no mystery surrounding the demand of a ¼” flat washer. The comparative reliability of supply in that industry supports a stability in demand that is all too absent in the furniture industry. The entire supply chain – from customer to manufacturer – seems to accept the unpredictability of supply. A customer shopping for an end table may be disappointed to leave a store empty handed, but she probably entered with low expectations. If her local grocery failed to supply her need for a household staple, she would be outraged to leave empty-handed – largely because the US food industry has raised our expectations of its ability to meet our demands. Why shouldn’t our customers have the same expectations?

A simple idea

I recently read yet another manifesto that claimed to have found a simple way of mathematically organizing the supply chain to achieve the holy grail of high turns, exponential sales growth and lean inventory. On paper, this stuff always looks promising, but it’s like a diet fad – not terribly long after you implement the system, you’re searching for the new algorithm that will take you to the promised land. And yet, I’m one of those people who design those systems. Why is what we’re doing so elusive?

Maybe the objective – to organize the supply chain along a serialized set of transactions informed by mathematical projections – contains the seed of its own failure. The supply world is organic and unpredictable. Engineers see this as the challenge: let’s impose order and predictability! This works, to a degree – defined by the constraints you imposed when you fit the messy world into a statistical model. And then you become frustrated by the limitations you imposed.

I’m thinking now that a social model is more informative than a mechanical model. At a real basic level, we’re talking about people, not systems. People buy stuff – even when they follow departmental guidelines, they’re people: emotional, intuitive, cognitive, yet likely to make mistakes. In the supply world, there are individuals and groups of people interacting sometimes physically, sometimes virtually. All of these people are following some idea of a process – whether a customer or a truck driver or a production line supervisor – while they act in a personal and very individual way. Attempting to mechanize their actions and decisions is futile. Success for any of them is simply that the outcome of what they did was good: the customer found what she was looking for (at an acceptable price); the driver arrives on time without mishap; the supervisor’s line production and employee morale are high. So how can all of them be more successful?

You don’t have to over-think this one. We’re social animals. We naturally create organizing principles for living with each other. We communicate, tell stories, teach, learn, entertain. We share (sometimes) and take (sometimes). We create ideas about what we experience, and then change those ideas when experience changes our thinking. We make tools. When the tools break, we make new ones – and you have to count on the tool breaking; it will.

So how do I …… (forecast demand… plan resources….manage an assortment.. fill in your own need…)? Let’s not over-think it, or try to find the one super concept that will solve the problem for everyone, forever. Before we make a tool, let’s use those social attributes (communicate, tell stories, teach, learn, entertain) and tap into the combined experience of the people in our social group. When we make the tool, we’ll accept that it’s just one way of solving the problem. We also have to accept that not everyone in the group will be able to use the tool expertly every time, so the tool’s design has to cater for that. Then, be prepared to keep making it anew.

I think it’s that simple. And that complex.

Checkpoint or Checkmate?

At a meeting yesterday in our DC, I heard strong interest in establishing a system of quality inspection checkpoints so that we can better understand where quality defects occur. I was pleased and yet a bit surprised: quality control is classic give-and-take. Yeah, it’s great to have it, but it requires real effort to make it happen. It also takes a huge leap of faith to believe that the extra steps you are taking up-front will result in a productivity savings down the road. Typically, we all become so involved in what we have to accomplish right now that it’s hard to believe it’s worthwhile to produce somewhat fewer transactions so that we can spend more time making sure that each transaction is of good quality.

It comes down to what we value: quantity or quality? Is this a case of win/lose (checkmate), or is it possible to achieve both? What would a work environment look like in which we did achieve both?

Hands-free scanning

We handle large, often unwieldy cartons and SKUs in our warehousing operations. Capturing the receipt of an item requires at least 6 separate scans (item code, PO, package quantity, scan to cart, scan putaway of item to location) – which in our business means taking the hands off the product, picking up the RF scanner, locating and then scanning the appropriate bar code. And inevitably, you do this so many thousands of times and inevitably you drop the gun or leave it sitting where it shouldn’t be – and thus the guns are on a constant repair cycle.

I’ve been thinking lately about this problem – how could we reduce or (dare we dream?) eliminate scanning while retaining tight control of our inventory? What if we moved away from the data-limiting zebra-stripe barcodes, to the more data-rich 2D (QR) codes? An entire ASN line could be encoded in this expanded data symbology, and with one scan we could verify and receive an item.

Recent breakthroughs in voice technology (see Vangard) make me wonder about receiving by voice – completely hands free! The receiving label could display an ID code for the ASN line: using voice tags and simple codes the operator could receive the item, instruct the system what to do with discrepancies using specific commands, and putaway the item into the bin location. No juggling RF guns and product, no equipment to be damaged, and potentially we can increase accuracy since the operator’s mind is focused on the item and its location – not interpreting data on a label or on a tiny screen.

A similar solution for picking isn’t as obvious (yet) – but just the thought of streamlining the receiving and putaway process while increasing accuracy is exciting.

Why is it so slow? Why is it so expensive?

I’m working with operations and IT staff on an EDI project. Hauling through the standards, I experienced a flash back to the late 80’s and my first EDI implementation. Back then, IT projects moved at a snail’s pace by comparison, and I remember being grateful for a standards body to produce a standardized set of transactions that all industries could use. My expectations are much different now – development should be fast; it should be collaborative; existing standards and protocols should be readily available at no cost to encourage everyone to use them. It’s such a disappointment to see that the only changes have been for the worse.

And yet, I see more of a need for greater speed and collaboration and open standards as technology moves to the mobile platform. Mobile technology development isn’t just about the latest cool phone app… thinking differently about the way we develop collaborative interfaces is absolutely necessary for supply chains to reinvent themselves for the global market of the 21st century.