April 2011

Did you read recently, as I did, that the Office For National Statistics is compiling a "Happiness Index"? (This isn't a delayed April Fool - have a look here. )

Maybe I'm being cynical if I suggest that, at a time when disposable income is falling, the government would rather talk about quality than quantity. (Never mind the width, feel the quality!) But whatever the reason, this shows how far the concept of quality has spread.

You can see this in the growth of quality assurance as a discipline. In particular, the branded systems - such as Six Sigma - have gained a firm foothold. Although these systems were developed to control production, they've now spread well beyond that.

Of course, we should always seek to improve our logistics operations. But is this best done using systems developed for a different function? After all, these systems were designed to assure manufacturing tolerance. Does it make sense to use them if we replace that clear measurement with the vague definition: "process output required by the customer"?

Let's start positively; these systems have elements that we can definitely use.

First of all, the methodology. Here it is: define the problem; measure what's happening now; analyse the measurements you've collected; improve the current process based on your analysis; control the new process to maintain the new performance. (You'll see this methodology called DMAIC; but it's what you do that's important, not what you call it.)

This methodology is good practice, so it makes sense to use it. And in particular, look at the last point - control the new process. Too often, new processes fail simply because, once the change has been made, no-one follows it up.

Secondly, quality assurance programmes have a tool-box of techniques to help you diagnose the causes of failure. These techniques aren't unique to quality assurance, but you can find them all in one place. (If you want a list, just contact me)

But we have to be realistic. As you've seen, all these systems have developed from programmes which control manufacturing. That gives them two characteristics that don't fit logistics so well:

- they're designed for closed systems

- they're designed to manage a particular type of error

Manufacturing is usually a closed system. That is, all the key factors for quality are within the control of the organisation that's responsible for quality. Logistics isn't like that. Take customer delivery as one example. If your delivery vehicle leaves your depot late, that's an error that you can rectify - but if it's delayed on the way to your customer by an accident, there's nothing you can do to control that.

Control of production processes is based on two key assumptions about error. First of all, that error is distributed normally - for example, if you're machining a component, you're as likely to make it too large as too small. Secondly, that both types of error are equally bad. A component that's too large is as useless as one that's too small.

These assumptions make sense for manufacturing - but they don't always apply to logistics.

In many logistics processes, error isn't distributed normally. Think about customer deliveries again. When you compare due delivery time with actual delivery time, then you usually have much larger errors on the late side than you have (or even can have) on the early side.

Not only that, although delivering early is a variation from the standard, it's a variation that often doesn't matter - in which case it isn't an error.

This doesn't mean you shouldn't define a standard and measure your performance against it. But it does mean that many of the assumptions and techniques (including the concept of six standard deviations - or six sigma - itself) don't work for the logistics processes we have to manage.

So what's the verdict on implementing quality assurance programmes in logistics operations? Let me summarise it for you (if you want to know more, you can always contact me here - it's free and without obligation).

Take note of the methodology - define, measure, analyse, improve and - especially - control. Check out the techniques these programmes use, and pick out those ones which will help you.

But don't expect to transfer manufacturing quality systems directly to logistics. There are real differences between the two disciplines that you can't ignore.

<< Back