XYZ Analysis
XYZ analysis is a method that classifies items by how predictable their demand pattern is. X items have steady, easily planned consumption, Y items fluctuate for seasonal or trend reasons, and Z items are consumed irregularly and are hard to forecast.
XYZ analysis is a classification method that sorts items by how regularly, and therefore how predictably, their consumption behaves. It assigns each item to one of three classes: X items show largely constant demand with only minor fluctuations and can be forecast very accurately. Y items are subject to recognisable but limited fluctuations, for example due to season or trends, and are moderately plannable. Z items are consumed completely irregularly and sporadically; their demand is hard to predict.
Unlike the ABC analysis, which ranks items by their share of value, the XYZ analysis assesses only the consumption structure over time. It does not answer the question "How important is an item in terms of value?" but "How well can I forecast its demand?". The result is the basis for choosing the right procurement and stocking strategy for each item, from consumption-driven inventory for X items to demand-synchronised procurement for Z items.
At a glance
- Classifies items by demand predictability, not by value
- X = constant and easily planned, Y = fluctuating, Z = sporadic
- The calculation is based on the coefficient of variation of consumption
- Usually combined with the ABC analysis to form the ABC-XYZ matrix
- Drives stocking strategy, safety stock and the planning method
How XYZ analysis works
The basis of the XYZ analysis is an item's historical consumption over several periods, typically the monthly issues of a year. For each item, it measures how much the actual consumption fluctuates around its own mean. The smaller this fluctuation, the more reliably future demand can be derived from the past and the “better” the item is in the sense of the analysis.
The key figure is the coefficient of variation: the standard deviation of period consumption divided by the mean period consumption, usually expressed as a percentage. A low coefficient of variation means even consumption (X), a high one means widely scattered consumption (Z). The company sets the boundaries between the classes itself; common values are below 10 percent for X, 10 to 25 percent for Y and above 25 percent for Z. These thresholds are not a standard but are adapted to the industry and the product range.
The three classes in detail
X items have almost unchanging consumption, such as packaging material or a standard item with constant sales. Their demand can be forecast very accurately with simple forecasting methods. Y items follow recognisable patterns with more pronounced swings, for example seasonal goods such as barbecue accessories or Christmas items; their pattern is still manageable with seasonal forecasting models. Z items, finally, are consumed with no recognisable pattern, such as spare parts or slow-moving items with occasional large orders. For them a statistical forecast is hardly worthwhile, which is why they are often procured only when there is concrete demand.
XYZ analysis combined with ABC analysis
The XYZ analysis reveals its full value only in combination with the ABC analysis. The two methods look at different dimensions: the ABC analysis ranks items by their share of value or revenue (A = high value, C = low value), the XYZ analysis by the predictability of demand. Overlaying both produces the ABC-XYZ matrix with nine fields, assigning each item a combination such as AX, BY or CZ.
Clear recommendations for action can be derived from this matrix. AX items are valuable and easily planned; here a lean, consumption-driven stocking approach with low safety stock pays off, because the capital-tie-up risk is high but the forecasting risk is low. CZ items are low in value and irregular; they are stocked pragmatically with simple rules or not at all. AZ items are considered critical: high value combined with poor plannability, which calls for especially careful, often demand-synchronised procurement.
Why XYZ analysis matters
The benefit of the XYZ analysis is that it makes a product range steerable in a differentiated way instead of treating all items alike. A company with thousands of line items cannot plan every item individually. The classification provides the decision basis for which items tolerate automatic, stock-oriented replenishment and which must be procured closely to actual demand. This lowers capital tie-up and inventory costs for plannable items while at the same time reducing the risk of shortages for critical items.
Concretely, the class influences the level of safety stock and the planning method. X items get by with low safety stock because their consumption is predictable. Z items require either a high buffer, if being unable to deliver would be costly, or order-related procurement with no significant inventory. In this way the XYZ analysis links range analysis directly to operational decisions in planning and inventory management. Different target service levels are also often defined per class: a moderate target service level is enough for easily planned X items, while for important but poorly forecastable items a deliberately higher service level determines the necessary safety buffer.
Limits and pitfalls
The XYZ analysis is only as good as the underlying consumption data. With a short data history, new items or one-off outliers, the coefficient of variation yields misleading values. The classification is also a snapshot: if demand behaviour shifts, the analysis has to be recalculated regularly, for example quarterly. And it says nothing about the cause of the fluctuation; whether a Z pattern stems from genuine sporadic demand or from data errors has to be judged on the merits.
XYZ analysis in the ERP system
In ERP and inventory management systems, the XYZ analysis is an evaluation based on transaction data. The system reads the consumption or issue postings per item over a defined period, calculates the mean and the dispersion and automatically assigns an X, Y or Z class based on stored thresholds. This class is often saved as an attribute on the item master and can be linked with the ABC class for a combined ABC-XYZ assessment.
The real added value arises when the classification is not only evaluated but fed back into planning. Advanced systems automatically select the appropriate method depending on the XYZ class: for X items a consumption-driven reorder via reorder point and order point method, for Z items demand-driven or manual planning. The prerequisite is clean, complete inventory management, because only reliable transaction data lets the analysis deliver robust classes. Some systems ship the evaluation as a standard function, while in others it is handled via BI tools or exports.
Example
Example: office supplies wholesaler
An office supplies wholesaler carries around 8,000 items and evaluates the monthly issues of the past year in its ERP. Copy paper sells almost identically month after month; the coefficient of variation is 6 percent and the item is classified as X. A set of Christmas greeting cards shows a clear seasonal peak in the fourth quarter and a coefficient of 19 percent; it falls into class Y. A rarely requested heavy-duty shredder is ordered only in isolated months with high quantities, a coefficient above 40 percent, so Z.
Combined with the ABC analysis, the copy paper ends up in class AX: high revenue, very easily planned. The planner switches it to automatic, consumption-driven replenishment with low safety stock and thus lowers capital tie-up. The shredder, as a BZ item, moves to procurement-synchronised ordering: it is ordered only once a concrete customer order exists, instead of being kept permanently in stock.
Frequently asked questions
Matching ERP systems
Related services
Questions about XYZ Analysis in your ERP project?
We advise vendor-neutrally – and implement it ourselves on request.