Skip to content
Controllingblog

Home❧Software & Data❧Data warehouse explained simply

Data warehouse explained simply

One truth instead of five exports: what a data warehouse is, how it works and when it is worth it for mid-sized companies too.

Software & Data7 min readUpdated 14 Nov 2026
Illustration: laptop with dashboard, database and cloud +12% 4.2 M 98%

In many companies every analysis starts with the same exercise: export data from accounting, export data from the ERP, export data from the CRM – and combine it all in Excel. Result: every department has slightly different numbers, and meetings discuss the numbers instead of the decisions. A data warehouse addresses exactly this.

What is a data warehouse?

A data warehouse (DWH) is a central database in which data from various source systems is collected, cleaned and prepared for analysis. It is optimised for reading and analysing, not for daily entry of transactions. The goal is a “single source of truth”: one reliable data basis for all reports.

The ETL process

  1. Extract: data is read from the source systems – financial accounting, ERP, payroll, CRM.
  2. Transform: it is cleaned and standardised: same customer numbers, same account logic, currencies converted, duplicates removed.
  3. Load: the prepared data is loaded into the warehouse – typically every night.

Modern approaches partly reverse the order (ELT): raw data is loaded first and only transformed in the warehouse. For controlling this changes little – what matters is that the preparation rules are central and traceable.

Facts and dimensions

Data in the warehouse is usually organised in a star schema:

  • Facts are the numbers: revenue amounts, quantities, hours, costs.
  • Dimensions describe the numbers: when (date), what (product), who (customer), where (cost centre, entity).

This separation makes analyses flexible. Whether you want revenue by month, customer or product group – it is always the same fact table, just filtered differently.

Data warehouse, data lake, database – what is what?

Is it worth it for mid-sized companies?

A data warehouse doesn't have to be a big project. For many mid-sized companies a lean database in the cloud or on their own server is enough, filled from three to five sources, with Power BI or Excel as the front end. It is worth it if:

  • the same analyses are regularly assembled from several systems,
  • numbers don't match between departments,
  • historical data would be lost when switching systems,
  • several entities are to be analysed together.
Practical tipStart small: one source system (usually financial accounting), one report (monthly reporting), one clean data model. When that works, add the next sources. Big projects that want everything at once fail strikingly often.

The role of controlling

A data warehouse is not purely an IT project. Controlling contributes the business logic: which KPIs, which definitions, which hierarchies (income statement structure, cost centre tree, product groups)? Without this work you get a technically perfect system that nobody understands.

Conclusion

A data warehouse is the foundation for reliable reports. It saves recurring manual work, creates a common numbers basis and preserves history. Start small, define the business logic cleanly and expand step by step – then it is very feasible for mid-sized companies too.

Further links

Frequently asked questions

What does ETL mean?

Extract, Transform, Load: data is read from source systems, cleaned and standardised, and then loaded into the data warehouse.

What is the difference between a data warehouse and a data lake?

A data warehouse stores cleaned, structured data for analysis. A data lake stores raw data in any format for later analysis.

Do I need my own IT department for a data warehouse?

Not necessarily. Lean solutions can be built with external support. What matters is that controlling owns the business logic.