If the ABC inventory management approach is a good idea, why not ABCD or ABCDEF…and so on? If three classes of inventory help it perform better, why not more classes? What’s the right number? How about the “ultimate ABC” – a class for every item?

While desirable, this is not practical (usually not even possible) using the ABC approach. However, a new method, Optimized Inventory Management, does exactly that – have a class for every item and treat every item on its own merits. The optimized approach has the added benefit of providing better service and lower cost than any other approach.

The ABC approach is widely used…very widely used. In a reader poll taken by a leading inventory management newsletter, the ABC approach was ranked number one as an inventory management technique. Fully 36.7% of the newsletter’s readers responded to a recent survey indicating ABC as the preferred technique.

The Ultimate ABC
It seems that the ABC approach is such an well-accepted idea that we wonder why others have not extended it to its logical conclusion. If having three classes, As, Bs and Cs, are a sensible idea, then why not have four or five classes? If five works well, why not more?

After all, the items in the A class are different from each other – different forecasts, different costs, lead times, and so on. Having three classes (or any number less than the number of items) creates groups of different items that are still treated the same way. The whole idea behind the ABC approach is to treat different items differently!

What’s been gained by grouping all the items into a few classes? With three classes we still need to figure out how to manage every item in the class. With three classes all we’ve done is create three problems rather than one!

If you believe the ABC idea, the only logical conclusion to come to seems to be the ultimate ABC – a class for every item.

The ABC Problem
But there’s a problem with this…a big problem.

In the ABC approach the user establishes classes of items and then manages the classes. This is done with “control parameters” that are set to different levels for each class. Different “target service levels”, different “weeks on hand” or whatever control parameters are used in the particular system. But here’s the problem. With additional classes we have to figure out what values to set their control parameters to. Rather than three different “weeks on hand” values, we’d need four values – one for each class.

This is an even bigger problem when we realize that just setting the three values originally was hard enough.

In reality there’s no fundamentally sound way to set the control parameters of the three classes in the first place! Some say, “well, the As are more important than the Bs so we set the target service level of the As to 96% and the Bs to 94%”. But these values are arbitrary. There’s no rational reason for them. Why not 97 and 93% or any other values?

Yet these values have dramatic impacts on both the service provided by the inventory and the capital it requires. And there’s no good way to set these values even for a few number of classes! This problem is a built-in incentive to have just a few classes – even one – negating the whole class idea. How many of us have seen ABC systems with the control parameters of every class set to the same values?

This arbitrary nature of the ABC approach even applies to the determination of the class itself. There’s no rational way to say where the As stop and the Bs begin. (Using the 80/20 rule doesn’t make it less arbitrary.) Why should the last A item be managed very differently than the first B item?

This problem is the principle reason the ABC approach is not more widely used. Inventory managers just don’t have a sound way to establish classes or of setting the parameters of each class. Most employ values they’ve heard other companies use. Almost everyone uses trial and error. Then, in practice, managers react to pressures that, at some times, lower certain values in response to cash needs and, at another times, raise some in response to poor service.

The Solution
The solution to this problem is, actually, easy. It’s to not require the inventory manager to set the control parameter values of each class (let alone each item). Have the computer do it! Have the computer figure out how to manage each class (item) so as to achieve a goal. With such a capability you can get rid of the arbitrary class structure – you can have as many classes as possible. One for every item – no problem. Then you can manage every item according to it’s own characteristics – it’s own forecast, cost, lead time, and so on – the ultimate ABC.

Why should inventory managers have to tweak parameters for each class? They should strive for a more strategic goal. High turns, increased service level, lower cost or some such goal; retaining customers, more sales or market share. Setting the control parameters of the As to 96% and the Bs to 94% is very far removed from such goals. Moreover, it may actually negate them. If you don’t know the 96 and the 94% values support the goal, chances are they don’t. If you can’t show your classification of As and Bs supports such a goal, it probably doesn’t.

The Optimized Approach
The Optimized approach to managing inventory has a specific goal. It maximizes the service provided per dollar invested in it. With this goal, the inventory manager can perform one of two possible optimizations. He or she can establish a service target, 97% say, then achieve it with the lowest possible capital investment in the inventory. On the other hand, the manager can establish a capital level – $30 million, for example. Then the Optimized approach will use that capital and achieve the highest possible service level with it. We all live with budget constraints – what a practical capability!

What good is the “service per dollar” goal? Maximizing service per dollar guarantees service and capital improvements over other approaches – all other approaches. This method is guaranteed to increase service and decrease capital over any other inventory management method. That’s what’s meant by “optimal”.

Maximizing service per dollar invested is very similar to maximizing turns. It has the added advantage that it’s the goal that keeps customers coming back. It maximizes the chance (for a given investment) that customers will find what they want, preventing them from going to a competitor. In this way, the goal of maximizing service per dollar insures the long-term success of the business. It beats the competition by garnering all the customers – theirs too.

How does the Optimized approach do that? It’s pretty straight forward. Have the computer search through all possible stocking positions and pick the best ones – the “optimal” ones. Search through stocking position of 0, 1, 2, 3, … and so on for every item. In this way every conceivable stocking position is evaluated at the lowest possible level of detail. Then the best stocking positions for a given capital level (or for a given service level) can be identified.

Realizing this, it’s easy to see how the Optimized approach treats every item as its own class. In this way, the Optimized approach accomplishes three benefits. It eliminates the ABC problem, it achieves the “ultimate ABC” – treating every item on its own merits, and it provides better service and lower capital than any other approach.

The biggest benefit of the Optimized approach, however, is not solving the “ABC problem”. Far more important is achieving high service levels and low costs. It beats other methods – all other methods – at the two things that matter most in inventory management – service and cost.