Data extraction is used in a variety of data management use cases and initiatives. Here is an example use case to consider:

A large ecommerce company wants to optimize its customer retention strategy. They extract customer data from their sales database, including purchase history, demographics, and customer service interactions. 

Although not part of the data extraction process itself, this extracted data is then cleaned, standardized, and integrated with other data sources, such as website analytics and social media interactions. 

By analyzing this comprehensive dataset, the company can identify trends, segment customers, and implement targeted marketing campaigns to increase customer loyalty and drive sales.

Yes, data can be extracted outside of the ETL or ELT processes. However, extracted data by itself – without the subsequent transformation and loading steps – is often raw and unstructured. This greatly limits its usefulness for analysis and decision-making.

In some cases, data is extracted (without transformation or loading) for archival or compliance purposes. However, its full potential is realized when it's integrated into the broader data pipeline. 

While both terms may seem interchangeable, these are two different processes with different outputs and purposes.

Data extraction focuses on retrieving raw data from disparate sources – without inherently offering any sort of analysis, intelligence, or insights. In other words, the end product of data extraction is raw data. 

Data mining, on the other hand, goes one step beyond data extraction. Using extracted data, data mining enables a certain degree of analysis, trend reporting, and insights. The end product of data mining is actionable intelligence.