Syntha Lab RU

ABC-XYZ matrix: a SKU’s contribution to sales and demand stability

From a table of SKU sales over several periods splits them into classes A, B, C by contribution to sales and X, Y, Z by demand stability (the coefficient of variation), shows the matrix of sales shares and the number of SKUs and suggests a stock policy for each SKU.

Assortment and sizes For whom: Category managers, buyers and stock planners when revising the stock policy.

Calculate on your data →

At a glance

What question does it answer?
Which SKUs bring sales, which of them are stable, and what stock policy to assign to them.
What data is needed?
SKU sales over equal periods, one metric.
How to read the result?
Example on the default values (12 illustrative SKUs, 6 periods): class A gives 84.7% of sales with 58.3% of SKUs, AX articles 4 with a sales share of 58%, with unstable demand 4, candidates to drop CZ 1. Example on the default values: SKUs in the calculation: 12; Share of class A SKUs: 58.3%; Share of class A sales: 84.7%; AX SKUs: 4. The figures are illustrative: enter your own to see your result.

Why and how to reason

Not all SKUs are equally important, and not all are equally predictable. So they are classified along two axes. ABC answers “how much money does it bring”: SKUs in descending sales, class A are those that give the first 80% of cumulative sales, B up to 95%, the rest C. XYZ answers “how stable is demand”: the coefficient of variation (standard deviation ÷ mean) up to 30% is X, up to 60% Y, above Z. The basic jacket sells 180, 190, 175, 185, 195, 180: variation 4%, class X. The seasonal dress 20, 40, 110, 180, 230, 60: variation 78%, class Z. The 3 × 3 matrix gives a policy. AX: auto-reorder and high service. CZ: a candidate for removal. Here there are 12 SKUs, 58% of them class A give 85% of sales, AX are 4 items (58% of sales), unstable Z are four.

When to use it

Reach for it when setting stock policy, choosing which items to reorder automatically and where to keep safety stock, and when cleaning the assortment. Do not use it on short series (fewer than 6 periods) and on novelties: variation without history is random. Assumptions: the 80/95 and 30/60 boundaries are yours and can change, sales are taken in one valuation and for comparable periods.

How to read the result

First look at the matrix of sales shares: how much money is in each cell. AX and AY hold the business: they need high service and an accurate forecast. AZ are the hardest: much money and unstable demand, a more frequent order and safety stock are needed. Candidates for removal CZ (here 1, their sales share is small) should be checked before removal: perhaps it is a novelty or an item for a set. If the sales share of class C is 4% but there are many SKUs, you pay for complexity: consider cutting. For each SKU the calculation suggests a policy, but the decision remains yours.

Common mistakes

The usual mistakes: counting variation of a seasonal item over a year as instability although seasonality is predictable (use seasonal indices), taking ABC by quantity rather than money or margin, including novelties without history. Revisit classes quarterly, not once a year. Do not leave classes without decisions: every cell of the matrix should lead to a concrete ordering rule, otherwise the classification remains a report.

What data is needed?

  • SKU sales over equal periods, one metric.
  • The A, B and X, Y limits: yours, there is no single standard.
  • For seasonal SKUs the season’s periods.

Inputs

FieldUnitValue in the exampleMeaning
Class A limit% of cumulative sales80A SKU is in A while the cumulative share of sales before it is below this limit. Your choice.
Class B limit% of cumulative sales95C follows. Must be above the A limit.
Class X limit% variation30The coefficient of variation not above this value: demand is stable. There is no single standard; for fashion the limits are wider than for basics.
Class Y limit% variation60Above this is class Z: demand is unstable. Must be above the X limit.

How it is calculated

StepFormula
ABC classA while the cumulative share before the article is below the A limit; then B up to the B limit; otherwise C
Variationstandard deviation across periods ÷ mean, in percent (sample standard deviation, divisor n − 1)
XYZ classX when variation is not above the X limit; Y not above the Y limit; otherwise Z
Policyby the 3 × 3 table: from automatic reorder for AX to dropping for CZ

Worked numbers

Default values from the app; in the interactive version you enter your own.

Articles in the calculation
12
Share of class A articles
58.3 %
Share of class A sales
84.7 %
AX articles
4Large and stable: the base of the range.
Share of AX sales
58 %
Articles with unstable demand (Z)
4
Share of class C sales
4.3 %
Candidates to drop (CZ)
1
Share of sales by class, %
Heat map: 3×3; min 0, max 58XXYYZZAA58A / X: 586.5A / Y: 6.520.3A / Z: 20.3BB7.7B / X: 7.70B / Y: 03.3B / Z: 3.3CC2.7C / X: 2.70C / Y: 01.6C / Z: 1.6Demand stability: X stable, Z unstable
  • 0 … 58 %
Data as a table
XYZ
A586.520.3
B7.703.3
C2.701.6
Articles: demand variation and sales share
Scatter plot: 12 points05101520020406080100Basic jacket: 4; 19.3Basic jacketStraight jeans: 4.42; 16.8Straight jeansWhite T-shirt: 4.49; 12.4White T-shirtSeasonal dress: 78.1; 11.2Seasonal dressBasic sneakers: 7.88; 9.47Basic sneakersMid-season coat: 86.6; 9.09Mid-season coatShirt: 44.3; 6.47ShirtSocks (set): 6.08; 4.28Socks (set)Scarf: 28.8; 3.41Hat: 91.9; 3.32HatBag: 11.1; 2.71BagBelt: 95.5; 1.56BeltDemand variation, %↑ Share of sales, %
Data as a table
Demand variation, %Share of sales, %
419.31Basic jacket
4.4216.78Straight jeans
4.4912.41White T-shirt
78.1411.19Seasonal dress
7.889.47Basic sneakers
86.559.09Mid-season coat
44.316.47Shirt
6.084.28Socks (set)
28.783.41Scarf
91.893.32Hat
11.062.71Bag
95.481.56Belt

Articles by descending sales

ArticleSalesShare, %Cumulative, %ABCVariation, %XYZClassPolicy
Basic jacket1,10519.319.3A4XAXKeep in stock always, automatic reorder
Straight jeans96016.836.1A4.4XAXKeep in stock always, automatic reorder
White T-shirt71012.448.5A4.5XAXKeep in stock always, automatic reorder
Seasonal dress64011.259.7A78.1ZAZOrder in small parts, watch the stock
Basic sneakers5429.569.2A7.9XAXKeep in stock always, automatic reorder
Mid-season coat5209.178.3A86.6ZAZOrder in small parts, watch the stock
Shirt3706.584.7A44.3YAYHold with a cushion, reorder by forecast
Socks (set)2454.389B6.1XBXAutomatic reorder with a stock cap
Scarf1953.492.4B28.8XBXAutomatic reorder with a stock cap
Hat1903.395.7B91.9ZBZOrder to demand, do not pile up
Bag1552.798.4C11.1XCXMinimal stock, rare orders
Belt891.6100C95.5ZCZA candidate to drop or to make to order

Classes are relative to your range and your limits: there is no single standard for the limits. Z for a seasonal article often means seasonality and not weak demand; compare periods of one season. For footwear and children’s count at the model and size-core level: demand by single size is jagged. For accessories check dropping an article against margin and stock.

Nuances and limits

  • Calling a seasonal SKU unstable over a year: compare periods of one season.
  • Dropping CZ without checking margin and role in the look: an accessory may be needed.
  • Taking zeros for no demand when the item was out of stock: an empty cell means “no data”.

Terms used

Calculate on your data →