Intermittent spare-parts forecasting, from evidence to accountable baseline.
Sparse demand is not ordinary demand with extra zeroes. Audit availability, describe the pattern, compare methods, test unseen periods, and measure whether overrides actually help.
Stockout-Censored Demand Auditor
Find periods where observed sales are only a lower bound on requests.
02 · ClassifyIntermittent Demand Pattern Classifier
Calculate ADI and CV² without hiding zero-demand periods.
03 · BaselineCroston, SBA & TSB Forecast Calculator
Compare transparent level forecasts and a simple mean baseline.
04 · ValidateIntermittent Demand Backtest Comparator
Rank methods on rolling one-step-ahead errors without future leakage.
05 · GovernForecast Value Added Auditor
Measure whether human overrides improved error and bias.
Five different decisions, one controlled workflow.
The tools do not split one score across pages. The censoring audit decides whether sales history represents demand; classification describes the demand structure; forecasting creates candidate baselines; backtesting evaluates unseen performance; FVA audits a later human process after actuals exist.
Do not convert a forecast directly into a stocking verdict.
Intermittent-demand methods estimate a demand level. Stocking also needs lead-time uncertainty, consequence, repairability, installed base, substitutions, MOQ, shelf life, budget, and local service policy. Continue with the existing MRO planning tools only after the demand evidence is fit for use.