Detailed definition
Forecasting in BI estimates what may happen next for a business metric. It can use statistical methods, machine learning, business assumptions, or simpler trend logic depending on the use case and data quality.
The BI context matters because forecasts need to be understandable and actionable. A useful forecast should connect to the metric definition, the historical data behind it, the uncertainty around it, and the decisions it supports.
Why it matters
Teams need forward-looking analytics, not just historical reporting. Forecasting helps leaders plan inventory, staffing, pipeline coverage, budget, customer risk, and operational capacity.
Forecasting also becomes more valuable when paired with alerts and what-if analysis. Teams can monitor whether actuals are diverging from expectations and test possible responses.
How it works
The system selects a metric, time grain, historical window, model or assumption set, and forecast horizon. It then calculates expected future values and presents them with relevant context.
In governed BI, the metric being forecast should come from the semantic layer. That keeps the forecast aligned with the same definition used in dashboards and reports.
Practical examples
- Forecast weekly revenue for the next quarter.
- Estimate inventory demand by region and category.
- Predict churn risk by customer segment.
- Compare actual sales with forecasted sales in a scheduled business review.
Common pitfalls
- Forecasts are not facts. They should be labelled clearly and reviewed against actuals.
- Forecasting cannot fix poor metric definitions or missing historical data.
- A forecast without business context may be hard to act on.