ABC analysis sorts your products into three groups by how much of your annual spend or sales value each one accounts for, so that management effort goes where it pays. A small number of products always dominate the value; a large number contribute very little.
The purpose is to stop treating every product the same. Applying identical reorder rules, service levels and counting frequency across a whole range means spending too much attention on items that do not matter and too little on the ones that do.
The Pareto principle behind it
ABC analysis is an application of the Pareto principle — the observation that roughly 80% of outcomes come from roughly 20% of causes.
In inventory, the pattern is remarkably consistent: around 20% of your products typically account for around 80% of your annual consumption value. The exact split varies, and the shape almost never does. Run the analysis on any reasonably sized range and you will find a small group of products carrying most of the value, a middle group, and a long tail of items that individually amount to very little.
The figures are a description, not a rule. Do not force your data into an 80/20 split. Some businesses find 10% of products carrying 70% of value; others find 30% carrying 85%. The useful output is the shape of your own curve, and where the natural breaks fall in it.
The three classes
A items — high value, tight control. Typically the top 10–20% of products by annual consumption value, accounting for roughly 70–80% of it. These justify individual attention: accurate forecasting, carefully calculated reorder points and safety stock, frequent counting, and close supplier management. A stockout here is expensive and so is excess.
B items — moderate value, routine control. Typically the next 30% or so of products, accounting for roughly 15–25% of value. Managed by rules rather than by individual attention. Reviewed periodically, counted less often, with sensible automated replenishment.
C items — low value, simple control. Typically the remaining 50% or more of products, accounting for perhaps 5% of value. Managed with the least possible effort: generous order quantities to reduce ordering frequency, simple minimum-maximum or par levels, and infrequent counting. The cost of managing a C item carefully usually exceeds the value of managing it at all.
One frequent misunderstanding. C items are not unimportant. A cheap component that stops a production line is critical whatever its value — so criticality has to be applied as an override, and that is a known weakness of the method covered below.
How to do an ABC analysis
Step 1 — list every product with its annual usage in units and its unit cost.
Step 2 — calculate annual consumption value for each: annual units × unit cost. This is the figure the whole analysis sorts on, and using unit cost rather than selling price is the convention, because you are ranking where your money sits.
Step 3 — sort descending by annual consumption value.
Step 4 — calculate each item’s percentage of the total consumption value.
Step 5 — calculate the cumulative percentage running down the sorted list.
Step 6 — draw the class boundaries. Conventionally at roughly 80% cumulative for A and roughly 95% for B, with the remainder as C. Adjust to where your own curve breaks naturally rather than forcing the convention.
A worked calculation
A business stocks ten products. Here is the full analysis.
| Rank | Product | Annual units | Unit cost | Annual value | % of total | Cumulative % | Class |
|---|---|---|---|---|---|---|---|
| 1 | Product D | 1,200 | $145 | $174,000 | 44.1% | 44.1% | A |
| 2 | Product H | 3,400 | $38 | $129,200 | 32.7% | 76.8% | A |
| 3 | Product A | 950 | $32 | $30,400 | 7.7% | 84.5% | B |
| 4 | Product F | 2,800 | $8 | $22,400 | 5.7% | 90.2% | B |
| 5 | Product B | 6,500 | $2.40 | $15,600 | 4.0% | 94.2% | B |
| 6 | Product J | 400 | $28 | $11,200 | 2.8% | 97.0% | C |
| 7 | Product C | 1,800 | $3.10 | $5,580 | 1.4% | 98.4% | C |
| 8 | Product G | 900 | $4.50 | $4,050 | 1.0% | 99.5% | C |
| 9 | Product E | 5,000 | $0.30 | $1,500 | 0.4% | 99.9% | C |
| 10 | Product I | 260 | $2.00 | $520 | 0.1% | 100% | C |
| Total | $394,450 |
Reading the result. Two products — 20% of the range — account for 77% of the annual consumption value. Three more take it to 94%. The remaining five products, half the range, account for under 6%.
The point the table makes is about attention. Product E moves 5,000 units a year, more than any other line, and represents 0.4% of the value. It does not warrant a carefully calculated reorder point. Product D moves 1,200 units and represents 44% — a week’s delay on that line matters more than anything that could go wrong with the entire C group.
A note on the boundaries. The natural break here is after rank 2, where cumulative value jumps from 77% to 84.5%. Following convention rigidly would place Product A in the A class at the 80% line. Either is defensible; what matters is that the boundary is deliberate.
What to do with each class
| A items | B items | C items | |
|---|---|---|---|
| Forecasting | Individual, reviewed regularly | Rule-based | Simple averages or none |
| Reorder method | Calculated reorder point, reviewed quarterly | Automated reorder point | Min-max or par levels |
| Safety stock | Calculated from measured variability, high service level | Calculated, moderate service level | Generous and simple, or none on substitutable items |
| Order frequency | More frequent, smaller quantities | Moderate | Less frequent, larger quantities |
| Cycle counting | Weekly or monthly | Quarterly | Once or twice a year |
| Supplier management | Close relationships, measured performance, second sources | Routine | Minimal |
| Review | Individual product review | Periodic, by group | Exception only |
| Who decides | A named buyer | Rules | The system |
The logic behind the differences. On an A item, holding an extra week of stock is expensive and running out is worse, so precision pays. On a C item, the labour cost of calculating anything carefully exceeds the money at stake — so the right answer is a generous buffer, infrequent large orders, and leaving it alone.
Cycle counting frequency is the clearest application. Counting everything equally means insufficient attention on the products where an error costs most. Counting A items weekly, B items quarterly and C items annually concentrates effort where record accuracy actually affects the business, and it is the standard way cycle counting programmes are structured.
The limitations of ABC analysis
It ignores demand variability entirely. This is the significant one. Two A items with identical annual value need very different safety stock if one sells steadily and the other is erratic. ABC alone will treat them the same.
It ignores criticality. A cheap fastener that halts production is a C item by value and an A item by consequence. Criticality has to be applied as a manual override.
It ignores lead time. A product with a sixteen-week lead time needs more planning than one available next day, regardless of value.
It ignores margin. Annual consumption value is a cost measure. A high-volume, low-margin product can rank above a lower-volume product that contributes far more profit — which is why GMROI is a useful companion.
It is a snapshot. New products have no history, seasonal products distort annual figures, and a product’s class changes as demand shifts.
It can become self-fulfilling. Neglect the C items enough and they sell less, which confirms their classification.
The practical response is to treat ABC as the first cut rather than the whole answer — applying criticality and lead time as overrides, and adding a second dimension for variability.
Going further: XYZ and ABC-XYZ analysis
XYZ analysis classifies products by how predictable their demand is, rather than by value.
- X items — steady, predictable demand with low variation. Easy to forecast, so a modest buffer suffices.
- Y items — variable demand with a recognisable pattern, usually seasonal or trending. Forecastable with effort.
- Z items — erratic, irregular demand with no reliable pattern. Difficult or impossible to forecast, so the response is either a generous buffer or a deliberate decision to accept stockouts.
Variability is usually measured with the coefficient of variation — the standard deviation of demand divided by the mean — with thresholds set to suit your own data.
Combining the two gives nine categories, and this is where the method becomes genuinely actionable.
| X — predictable | Y — variable | Z — erratic | |
|---|---|---|---|
| A — high value | Tight control, minimal buffer, JIT works well | Careful forecasting, moderate buffer, review often | Highest attention: generous buffer or make-to-order |
| B — moderate | Automated replenishment, low buffer | Rule-based with seasonal adjustment | Larger buffer, periodic review |
| C — low value | Simple min-max, large orders | Min-max with a wide band | Hold a generous buffer or consider discontinuing |
The two cells that matter most. AX items are high-value and predictable — the ideal candidates for just-in-time and lean buffers, because you can rely on the forecast. AZ items are high-value and unpredictable, and they are where most of the difficult inventory decisions in any business actually sit: expensive to buffer, expensive to run out of, and impossible to forecast well.
CZ items are worth a specific look. Low value, erratic demand, and frequently the largest count of SKUs. These are the natural candidates for rationalisation, for supplier-held stock, or for a deliberate policy of ordering only when a customer asks.
How often to redo it
Annually as a minimum, and quarterly if your range changes quickly.
Also redo it after a significant range change, a major shift in costs, the addition of a new product line, or a change in demand pattern such as a category growing sharply.
Watch for products changing class. A product moving from C to A deserves attention it is not currently receiving. One moving from A to C is usually a decline that nobody has flagged, and it is often the first sign of a product worth discontinuing.
And keep the overrides current. Criticality and lead time flags need reviewing alongside the classification, since a supplier’s lead time changing can matter more than a product’s value class.
Where to go deeper
- Inventory management techniques — where ABC sits among the methods.
- Cycle counting — the clearest application of the classification.
- What is safety stock — sizing buffers per class.
- What is inventory optimisation — the wider discipline this feeds.
- What is GMROI — the margin dimension ABC ignores.
- What is deadstock — what the neglected C tail becomes.
The one-line takeaway
ABC analysis ranks products by annual consumption value so that control effort matches what is at stake — tight management on the small group carrying most of the value, minimal effort on the long tail. Its real weakness is ignoring demand variability, which is why combining it with XYZ analysis produces a far more useful picture of where your difficult decisions actually are.
Frequently asked questions
What is ABC analysis in inventory? A method of classifying products into three groups by annual consumption value, so that management effort, service levels and counting frequency match what each product is worth. A items carry most of the value and get tight control; C items carry little and are managed with minimal effort.
How do you calculate ABC analysis? List every product with its annual usage and unit cost, multiply the two to get annual consumption value, sort descending, calculate each item’s percentage of the total and the running cumulative percentage, then draw class boundaries — conventionally around 80% cumulative for A and 95% for B.
What are A, B and C items? A items are typically the top 10–20% of products by value, accounting for 70–80% of annual consumption value. B items are the next 30% or so, accounting for 15–25%. C items are the remaining half or more of the range, accounting for perhaps 5% of value.
What is the difference between ABC and XYZ analysis? ABC classifies by value; XYZ classifies by demand predictability — X is steady, Y is variable but patterned, Z is erratic. Combining them gives nine categories, and the combination is far more useful because two products of identical value need very different buffers if one sells steadily and the other does not.
How often should you redo ABC analysis? Annually at minimum, quarterly if your range changes quickly, and immediately after any significant change in range, costs or demand patterns. Pay particular attention to products changing class, since a C item moving to A is not getting the attention it now needs.
What are the limitations of ABC analysis? It ignores demand variability, criticality, lead time and margin, and it is a historical snapshot that disadvantages new products and distorts seasonal ones. The practical response is to treat it as a first cut, apply criticality and lead time as overrides, and add XYZ for variability.
Is a C item unimportant? No. C means low annual consumption value, not low consequence. A cheap component that halts production or a packaging item that stops dispatch is critical regardless of its value, which is why criticality must be applied as a manual override on top of the classification.
Warehouse Wise is a free, independent library of guides on inventory management, warehousing and fulfilment, written for the people actually doing the work — new warehouse and inventory staff, small and growing retail and ecommerce businesses, and anyone who holds stock and wants to understand it properly. Every guide explains the term, shows a worked example with real numbers where it helps, and stops. We are independent: we do not sell software or equipment, we take no commission on referrals, and our guides are not ranked by who pays us.