Tracking signal
Σ (forecast − actual) ÷ mean absolute error over the same periods; a drift shows when the modulus stays above 4 for long
- Formula
Σ (forecast − actual) ÷ mean absolute error over the same periods; a drift shows when the modulus stays above 4 for long- Unit
- ratio
- Aggregation rule
- Snapshot at a date
- Source system
- Plan and budget model; POS: receipts
- Owning role
- Analyst (BI)
- Frequency
- Monthly
Passport
- Grain
- SKU or category × accumulated periods
- Numerator
- Sum of (forecast − actual)
- Denominator
- Mean absolute error for the same periods
- At zero or no data
- Mean absolute error is 0: “—” (no errors).
- Limits
- Bias is visible when the absolute value stays above 4 for a long time. The periods in the sum and in the mean error are the same; few periods give noise.
- Check example
- Sum of (forecast − actual), units: 30; Sum of absolute errors, units: 40; Number of periods: 3.
a/(b/c). Result: 2.25 - Related metrics
Version 1.0 · valid from 2026-11-16 · checked 2026-11-16
Related pages
Definition
The cumulative forecast error in units of the mean absolute error: shows a systematic skew.
A value outside ±4 is often read as a skew; the threshold is a guide, replace with yours.