Legacy ETL process of taking data from OLTP systems like SAP to a data warehouse has been a mammoth task, both due to the complexity of transformation and the short time window available to make the updated data available before business start.  

The ELT (Extract Load Transform) paradigm is changing this. While it involves the same work as legacy ETL, it operates by moving raw data from source system to destination data warehouse first and then leveraging the power and scalability of the Datawarehouse (typically on cloud using MPP architecture and columnar storage) to do the compute intensive transformations in record time. This eliminates the need to have an oversized ETL tool with a staging server used just for few hours. It also makes real time data transfer seamless

As we have increasing real time Semi structured and unstructured data with schema on read use cases, ELT becomes more of a choice rather than ETL which has been primarily for structured data.