Inventory planning

ABC–XYZ Classifier

ABC alone tells you which items are worth money. XYZ tells you which are predictable. Crossing them gives you nine groups, each needing a genuinely different stocking policy — which is the difference between an inventory strategy and a spreadsheet.

Paste real data safely — it is parsed in your browser and never sent anywhere.
Paste your data

One SKU per row. First column is the SKU code, second is unit cost, then one column per period of demand (months, weeks — whatever you have, minimum three). Copy straight out of Excel, or paste comma-separated values. A header row is detected and skipped automatically.

How the two axes work

ABC ranks SKUs by period consumption value (unit cost × annual demand), then cuts the cumulative curve: the items making up the first 80% of value are A, the next 15% are B, the last 5% are C. In most catalogues, A is 10–20% of the SKUs and C is more than half of them.

XYZ ranks the same SKUs by how predictable their demand is, using the coefficient of variation — standard deviation divided by mean demand across your periods:

CV = σ(demand) ÷ mean(demand) X : CV < 0.5 steady, forecastable Y : 0.5 – 1.0 variable, some pattern Z : CV > 1.0 erratic, effectively unforecastable

What to do with each box

ClassCharacterPolicy
AXHigh value, predictableThe best items you have. Tight min-max, high service level, frequent small replenishment, supplier scheduling agreements. Worth automating.
AYHigh value, variableHigher safety stock than AX, reviewed monthly. Forecasting effort pays back here more than anywhere else.
AZHigh value, erraticThe dangerous box. Do not hold blanket cover — it is expensive and often still misses. Manage by exception: supplier agreements, consignment stock, or make-to-order.
BX / BYMid valueStandard min-max with a periodic review. Automate the policy and keep human attention for A and Z items.
BZMid value, erraticReview the business case for stocking at all. Often better served on a short lead time than from stock.
CXLow value, predictableEasiest win in the catalogue. Order in large batches infrequently, hold generous cover — it costs almost nothing and eliminates a whole class of stockouts.
CY / CZLow value, unpredictableThe long tail. Simple two-bin or min-max, no forecasting effort. Review annually for obsolescence — this is where dead stock accumulates quietly.

The most common surprise

Teams expect the problem items to be A items. They are usually C items. Because nobody pays attention to the tail, C-class stockouts scatter across many customer orders and destroy order fill rate while barely denting line fill. Fixing CX items — cheap, predictable, with carrying cost, expiry and criticality reviewed — is often the fastest service improvement available.

Common mistakes

  • Classifying on quantity instead of value. ABC is about money. A fast-moving cheap item is not an A item, and treating it as one wastes planner attention.
  • Using too few periods. Short histories can be unstable and miss seasonality. Use representative periods and investigate intermittent demand separately.
  • Never reclassifying. Items migrate between classes as products mature. Rerun quarterly — an A item that quietly became a C is exactly where obsolete stock hides.
  • Letting zero-demand periods inflate CV. A part that sold once in twelve months has an enormous CV and lands in Z. That is correct, but do not then try to forecast it — it needs a stocking decision, not a forecast.
  • Applying one service level across the whole catalogue. The entire point of this analysis is to stop doing that. Feed the class into the safety stock calculator and set service levels by box.

Get the free KPI starter pack

Blank and invented sample copies, with editable targets and a reporting-month selector. No email required.

Turn this result into an action

Check the assumptions with your team, compare a second scenario, and keep the result in your monthly review.

Explore the matching workbooks · Methods and limitations