Skip to main content

Classify the portfolio

Classify your SKUs in the importance × predictability matrix to decide where to focus collaborative effort and where to automate.

For the strategic logic behind this classification, see Strategic portfolio classification.

Before you begin​

  • You need access to the Forecast Classification view inside the Forecast module.
  • The classification is calculated automatically; no additional data entry is required.

Steps​

1. Open the Forecast Classification view​

From the Forecast module, access the Forecast Classification view. You can display it in grid mode (visual matrix) or table mode (list of SKUs with their assigned quadrant).

2. Read the matrix quadrants​

Each SKU falls into one of four quadrants based on its combination of importance (economic impact on the business) and predictability (how well the model can forecast it):

QuadrantImportancePredictabilityRecommended strategy
A–YHighHighProtect. The model works well and the SKU is critical. Minimal intervention; continuous monitoring.
A–ZHighLowIntensive collaboration. Critical SKU but hard to forecast. Requires expert manual adjustment.
B–YLowHighAutomate. Low economic impact and easy to forecast. Let the model work on its own.
B–ZLowLowTactical actions. Low impact and hard to forecast. Evaluate whether it is worth keeping the SKU.

3. Review the % of portfolio per quadrant​

The view shows the percentage of SKUs (or economic value) that falls in each quadrant. This indicator is the key to prioritizing your time.

Alert threshold: A–Z > 25%

If the A–Z quadrant exceeds 25% of the portfolio, the system flags it as an alert. This means you have too many critical SKUs with unreliable forecasts. Prioritize validation and collaborative adjustment in that quadrant before any other.

4. Act according to quadrant​

After reading the classification, define your plan for the cycle:

  1. A–Z: dedicate expert review time (Commercial + Marketing). Apply the 50/50 Rule before adjusting.
  2. A–Y: review quickly; intervene only if you have qualitative information the model does not have (phase-in, phase-out, events).
  3. B–Y: do not invest manual review time. The model is sufficiently accurate.
  4. B–Z: evaluate in parallel with the product team whether these SKUs should remain in the active portfolio.

5. Monitor movement between quadrants​

Compare the classification across closed versions to detect whether a SKU improved (moved from Z to Y) or worsened (moved from Y to Z) its predictability. An SKU that consistently improves can move to automation; one that consistently worsens needs attention.

See also​