ABC analysis of inventory: which items deserve the most attention

What ABC analysis is and why it works

The method applies the Pareto principle to inventory: in most ranges a small share of items accounts for a large share of the value. Typical thresholds are A for the items that make up the first 70–80% of annual consumption value, B for the next 15–25%, and C for the rest. In a typical range, A is 10–20% of the items and C is half or more. These are rules of thumb, not a law: your own data decides where the lines fall.

The point is not the letters. It is to spend counting time, forecasting effort and working capital where a mistake is expensive, and to run the long tail with simple, cheap rules.

Choose the measure first

ABC analysis ranks items by one number, so the number has to match the decision you want to make:

• Consumption value (quantity used or sold × unit cost): the classic measure for inventory control, because it shows where money is tied up and where errors cost most.
• Revenue or gross margin: better when the question is which products the business lives on.
• Number of order lines or picks: the right measure for warehouse layout, because it shows which items the pickers touch most often. Slotting by picks is covered in our guide to bin locations.

Do not mix measures in one ranking. If you need two views, run two analyses.

How to do it step by step

1. Choose the period, usually the last 12 months, so seasonal items are not over- or underrated.
2. Export for every item the quantity used or sold in that period and its unit cost.
3. Calculate the annual consumption value of each item: quantity × unit cost.
4. Sort the items from the highest value to the lowest.
5. Calculate each item's cumulative share of the total value.
6. Assign classes by your thresholds, for example A up to 80%, B up to 95%, C for the rest.
7. Review the exceptions by hand: new items without history, critical spare parts, items being phased out.
8. Write down the rules for each class and repeat the analysis every quarter or half year.

Worked example (illustrative numbers)

A distributor has ten items with a total annual consumption value of 100,000 EUR. Sorted from highest to lowest, with the cumulative share in brackets:

• P-101: 3,500 units × 12 EUR = 42,000 EUR (42%)
• P-102: 20,000 EUR (62%)
• P-103: 12,000 EUR (74%)
• P-104: 8,000 EUR (82%)
• P-105: 6,000 EUR (88%)
• P-106: 4,500 EUR (92.5%)
• P-107: 3,000 EUR (95.5%)
• P-108: 2,500 EUR (98%)
• P-109: 1,200 EUR (99.2%)
• P-110: 800 EUR (100%)

With the thresholds of 80% and 95%, the three items P-101 to P-103 form class A: 30% of the items and 74% of the value. P-104 to P-106 are class B with 18.5% of the value. The four items P-107 to P-110 are class C: 40% of the items, only 7.5% of the value. A counting error of 10% on P-101 is worth 4,200 EUR; the same error on P-110 is worth 80 EUR. That difference is the whole argument for treating them differently.

Give each class its own rules

• Class A: count often, for example monthly as a cycle count; review demand and reorder points regularly; calculate safety stock carefully; keep close contact with the suppliers.
• Class B: count quarterly; review parameters a few times a year; standard reorder rules.
• Class C: count once or twice a year; order larger quantities less often to save handling; check regularly for items that no longer sell and consider removing them from the range.

Many teams add an XYZ analysis on top, which classifies items by how steady their demand is. An AX item is valuable and predictable; a CZ item is cheap and erratic. The combination helps decide where forecasting is worth the effort.

How it works in OrgLines

The classification itself is a calculation on your sales or consumption data, and a spreadsheet is enough to run it. What matters is that the result changes how the stock is managed day to day. In OrgLines stock is counted per warehouse, and you can enter the cost of a product to see its margin, which gives you the unit cost the analysis needs. Products can carry your own custom attributes, so the class can be recorded on the product itself.

For the rules per class, a stocktake can cover one warehouse or all of them and reconciles the counted result with the system, and the inventory list suggests how much to reorder for each item. Purchase orders with goods receipt close the loop from the suggestion to the shelf.

Common mistakes

• Ranking by unit price instead of consumption value: an expensive item that sells once a year is not an A item.
• Using a period that is too short, so seasonal items jump between classes.
• Running the analysis once and never repeating it.
• Letting new items fall into C just because they have no history yet.
• Treating critical but cheap spare parts as unimportant: classify them by hand.
• Mixing value and pick frequency in one ranking.
• Producing the classes but not changing any counting or reordering rule.

Checklist

• The measure matches the decision: value, margin or picks.
• The period covers a full year.
• Unit costs are current and complete.
• Thresholds are written down and applied the same way every time.
• Exceptions were reviewed by a person.
• Each class has written counting and reordering rules.
• The next review date is in the calendar.

Try it free in OrgLines