A technical warehouse with five thousand items cannot manage everything with the same precision. An ABC analysis splits the assortment into groups so you can match attention and rules to each: strict where it matters, relaxed where it doesn't. This article explains how to build the analysis, what you do differently per class and why classifying on value alone is a classic trap. The broader framing is in spare parts management.
The Pareto principle
ABC analysis rests on the observation that a small share of items represents a large share of value. In many warehouses roughly twenty percent of items account for eighty percent of annual consumption value.
That skew is an opportunity. Focus your attention on that twenty percent and you control most of your stock value with a fraction of the effort.
The common split is:
- A items — around 20 percent of items, 70 to 80 percent of value
- B items — around 30 percent of items, 15 to 20 percent of value
- C items — around 50 percent of items, 5 to 10 percent of value
The exact boundaries are not law. Pick percentages that fit your own distribution and then keep them consistent.
Classify on value or on criticality?
This is where it most often goes wrong. Classic ABC analysis looks at annual consumption value: annual usage multiplied by unit price. That is a financial lens.
In a technical warehouse that is not the whole story. A two-euro gasket you rarely use is a C item by value, but if your plant's only furnace stands still without it, the financial weight of that part is enormous.
That is why many maintenance organisations work with two axes: value and criticality. A common addition is the VED classification — Vital, Essential, Desirable — where you determine the consequences of a stockout per item. Combine both and you get a matrix in which a cheap but vital part still receives the attention of an A item.
Which parts are vital follows from your maintenance analysis. An FMECA delivers exactly that information: which failure modes have the heaviest consequences, and which parts belong to them.
Building the analysis
Gather the data. Per item: consumption over the past twelve months and the current price. Twelve months absorbs seasonal effects; shorter periods distort the picture.
Calculate annual consumption value. Usage times price. Note: this is about consumption, not stock. An expensive part that has sat on the shelf for three years has a consumption value of zero.
Sort and accumulate. Order items from high to low and calculate the cumulative percentage. Where it crosses 80 percent lies the A/B boundary; around 95 percent the B/C boundary.
Correct for criticality. Review the C list for parts whose failure causes significant downtime and upgrade those.
Record the outcome. The classification belongs to the item in your system, not in a separate spreadsheet — otherwise it is untraceable within six months.
What you do differently per class
The analysis only becomes useful when different behaviour follows from it.
A items get tight control. Min/max is calculated with justification and revised periodically, orders are tracked, and you count them often — four times a year or more. You actively monitor lead times with the supplier.
B items get a lighter regime. Min/max based on a rule of thumb, two counts a year, ordering according to fixed agreements.
C items are mostly about not spending attention. Generous min/max so you rarely have to order, count once a year, and preferably an arrangement with the supplier that removes the administrative burden. A C item occasionally sitting surplus on the shelf is cheaper than the time spent managing it closely.
How to schedule those counting frequencies in practice is covered in cycle counting.
How often to revise
Once a year is enough for most organisations. If something substantial changes — a new production line, a machine taken out of service, a major price change — that is a reason to revise in between.
What you mainly want to avoid is a classification standing still while the plant moves. An item that was an A three years ago but belongs to a decommissioned machine holds onto attention and stock for no reason.
Frequently asked questions
Is ABC the same as a criticality analysis?
No. ABC is by default about financial weight, criticality about the consequences of failure. They complement each other and are often combined in technical warehouses, but they are two different questions.
How many items do I need before ABC is worthwhile?
The value starts to appear at a few hundred items. Below a hundred you can still manage the whole assortment with the same attention and the classification adds little.
What do I do with items without consumption history?
New items have no consumption yet and automatically fall into C. Assess those separately on criticality and expected use, and let them join the regular analysis after a year.
Can I use ABC to decide what to remove from stock?
Partly. A C item with no consumption in three years is a candidate for disposal, but first check that it isn't a vital spare for a machine that is still running.