What is a data warehouse?
Quick answer
A data warehouse is a central database designed for reporting and analysis. It holds cleaned, combined copies of data from many business systems (sales, billing, support, marketing) so BI tools can query one trusted place instead of each app separately.
Last updated
Updated · By Robert Breen
Why it matters
Operational systems are built to run the business: take an order, send an invoice. They are not built for questions like "revenue by customer segment over three years". Asking them is slow and may affect the live system.
A warehouse copies the data out (via ETL), keeps history and arranges it for fast summarizing. It also becomes the single place where the official definitions of numbers live.
A worked example
A 200-person services firm has data in Salesforce, QuickBooks, a time-tracking app and Zendesk. Each night their ETL loads all four into a Snowflake warehouse. A "client profitability" table joins revenue from QuickBooks with hours from time tracking for each client.
Before the warehouse, answering "which clients lose us money?" took an analyst a week of exports. Afterwards it is one Tableau dashboard, and it showed that 6 of their 80 clients were unprofitable once support hours were included.
Often confused with
Data warehouse vs database
A warehouse is a type of database, tuned for reading and summarizing large amounts of history rather than recording individual transactions quickly.
Data warehouse vs data lake
A data lake stores raw files of any shape cheaply. A warehouse stores structured, cleaned tables. Many companies use both.
Data warehouse vs a shared spreadsheet
For a small team, a well-kept shared spreadsheet can play the warehouse role. You need a real one when the data outgrows it or several systems must be combined.
Tips
- You do not need one to start BI; add a warehouse when you combine three or more systems.
- Common cloud options: Snowflake, Google BigQuery, Amazon Redshift, Azure Synapse and Databricks.
- Agree on table and column names early; renaming later breaks dashboards.
Related terms
Learn it in a free lesson
- What Is Business Intelligence? A Beginner’s Guide (about 8 minutes)
Questions about it
- What does a BI analyst do?: A BI analyst turns business questions into trusted numbers and dashboards. What the job involves day to day, the skills it needs and how…
Related in the AI glossary
Frequently asked questions
- Is Excel a data warehouse?
- No, though small teams sometimes use a master spreadsheet the same way. Warehouses handle far more data, many users and automated loads.
- What is a star schema?
- A common warehouse layout: one central "fact" table of transactions linked to "dimension" tables that describe customers, products and dates.
- Can Tableau connect to a data warehouse?
- Yes. Tableau Desktop connects to Snowflake, BigQuery, Redshift and others. Tableau Public is limited to files and a few web sources.
All business intelligence terms, A to Z · Free business intelligence lessons