Retyping a long passage from a PDF is a waste of time when the words are already in the file. Extracting text pulls the readable content into a plain .txt file you can paste into another document, feed into a translator, or analyse.
Most PDFs contain a hidden text layer — the actual characters that make the document selectable and searchable. Text extraction reads that layer directly. ColaPDF uses Mozilla's PDF.js engine to do this in your browser, so the document never has to be uploaded.
Scanned documents are different. A pure scan is just an image of a page, with no text layer, so there is nothing to extract until the file has been through Optical Character Recognition (OCR), which converts the picture of the text into real characters. If extraction returns nothing from a scanned file, that is why.
Extraction returns plain text, so complex layout — columns, tables, font styling — is not preserved. The goal is to recover the words quickly, not to reproduce the design. For a fast way to grab quotes, data or long passages, that is exactly what you want.