Glossary
Predictive analytics
Analytics that uses historical data and statistical models to estimate what is likely to happen next.
Also called: predictive modeling, forecasting analytics
Predictive analytics uses patterns in historical data to estimate future outcomes—next quarter's revenue, the likelihood a customer churns, or expected demand for a product. It sits after descriptive and diagnostic analytics in the maturity model: once an organization understands what happened and why, it can build models that project forward.
Underlying methods range from classical statistics, such as regression analysis and time series forecasting, to machine learning models trained on larger and more varied feature sets. Every predictive model is built and validated on historical data, so its outputs are estimates with uncertainty, not guarantees, and should be reported with a confidence range or error metric rather than a single number.
Predictive analytics is used across sales forecasting, demand planning, credit risk, and churn management, letting teams act ahead of an event rather than react to it. The core pitfall is overfitting: a model that fits historical noise as if it were signal will look accurate on past data and fail on new data. A second common failure is assuming the future will behave like the past when the underlying conditions—market, product, customer base—have genuinely shifted, which no amount of historical data can predict.
Last reviewed September 19, 2026