Business Intelligence
Also: BI, Business Intelligence, Informatique décisionnelle, Business analytics, Decision support
Technologies and processes that turn raw data into actionable insights via reporting, dashboards and analysis, so teams can decide based on facts rather than intuition.
What it is
Business Intelligence (BI) is the set of technologies, processes, and practices that collect raw data from across an organization, organize it, and present it as reporting, dashboards, and analysis. The goal is simple: turn scattered operational data into actionable insights that support day to day and strategic decisions.
BI is descriptive and diagnostic by nature. It answers what happened and why, in contrast to predictive or prescriptive analytics that estimate what will happen or recommend what to do.
Why it matters
- Single source of truth: everyone reads the same numbers, defined the same way.
- Speed: self service dashboards replace manual spreadsheet pulls.
- Accountability: KPIs are tracked, visible, and comparable over time.
- Cost control: fewer errors and less duplicated reporting effort.
Without BI, decisions rely on gut feeling or stale exports. With it, leaders monitor performance continuously and spot problems early.
How it is used in practice
A typical BI stack moves data through several stages:
- Sources: ERP, CRM, web analytics, payment systems, spreadsheets.
- Integration (ETL or ELT): extract, clean, transform, and load data.
- Storage: a data warehouse or lakehouse holds modeled, query ready tables.
- Semantic layer: shared definitions of metrics (for example, how revenue is counted).
- Consumption: dashboards, scheduled reports, and ad hoc queries.
Common deliverables include executive scorecards, sales pipeline views, financial variance reports, and operational monitoring screens.
Concrete worked example
A retailer wants to reduce stockouts. The BI team:
1. Pulls daily sales, inventory, and supplier lead time data into a warehouse.
2. Models a metric: days of stock remaining per product per store.
3. Builds a dashboard that flags items below a 7 day threshold in red.
4. Sets an automated alert to the regional manager each morning.
Result: a manager sees that 120 products in 8 stores will run out within a week, triggers early reorders, and cuts stockouts by 30 percent in one quarter. No data scientist required, just clean data and a well designed report.
Key points to remember
- BI is mostly about past and present, not prediction.
- Data quality and shared metric definitions matter more than tooling.
- Good BI is self service: business users answer their own questions.
See also
Frequently asked questions
What is business intelligence in simple terms?
Business intelligence (BI) is the set of technologies, processes and practices that collect raw data from across an organization, organize it, and present it as reporting, dashboards and analysis. Its purpose is to turn scattered operational data into decisions based on facts rather than intuition. BI is descriptive and diagnostic: it answers what happened and why.
What is the difference between business intelligence and predictive analytics?
BI looks at the past and the present: it explains what happened and why. Predictive analytics estimates what will happen, and prescriptive analytics recommends what to do about it. The same data warehouse often feeds both, but BI deliverables are reports and dashboards, not forecasts or models.
Who needs to understand business intelligence?
Data, marketing and finance leaders, plus any manager who reports on performance. A CDO owns the stack and metric definitions, a CMO reads acquisition and funnel dashboards, a CFO relies on variance reports. You do not need to build BI yourself to be accountable for the numbers it produces.
What are the stages of a business intelligence stack?
Five stages: sources (ERP, CRM, web analytics, payment systems, spreadsheets), integration through ETL or ELT to extract, clean, transform and load, storage in a data warehouse or lakehouse, a semantic layer that holds shared metric definitions, and consumption via dashboards, scheduled reports and ad hoc queries. Typical outputs are executive scorecards, sales pipeline views, financial variance reports and operational monitoring screens.
Can you give a concrete example of business intelligence solving a problem?
A retailer fighting stockouts loaded daily sales, inventory and supplier lead time data into a warehouse, modeled a metric called days of stock remaining per product per store, and built a dashboard flagging anything below a 7 day threshold in red, with a morning alert to regional managers. One manager saw 120 products across 8 stores about to run out within a week and triggered early reorders. Stockouts dropped 30 percent in a quarter, with no data scientist involved.