Your organization has multiple database that contain data related to customers, sales, HR and many other business functions. So why would you need a data warehouse?
The idea behind creating a data warehouse is to collect data from heterogeneous sources, aggregate those and create a normalized and a unified structure for the purpose of analysis, report generation and other business intelligence activities that look for patterns and insights within the data.
Why not just analyze the data and create reports from the source transactional databases, you may ask?
First of all, different databases and sources of data have different and incompatible schemas, which means each data source has its own unique structure. While modern analytics and reporting tools can connect to multiple data sources using a technique known as blending, there are multiple issues with using this approach. Firstly, creating reports from transactional databases are subject to inaccuracies because of the constant nature of changing data within transactional sources of data. Then, the dissimilar structure of each data source allows for inaccurate and disjointed reports.
Data warehouse resolve these problems by normalizing the data and creating stable and cleaned-up copies of data ready for reporting and intelligence gathering.
Data warehouse paired with ECTL solves this problem. The Extract, Clean, Transform, Load technique paves the way to provide a homogeneous structure from disparate sources of data, which in turn allows for analysis and interrogation of the data.