This works well with single-column documents, but not for multi-column ones. With those, Acrobat's optical character recognition function will often read from left to right across all the columns at once, and scramble the text from all of them together. I haven't yet learned of a way to isolate individual columns in the PDF and have them read one at a time.