I hadn't thought of the scan and formating angle.

With the right software tricks, you could probably get the PDF to print to a bitmap image file instead of a printer, then get the OCR software to read that file. Thus saving the analog/realworld step and getting a better image and a more accurate OCR.

One other question...

It's not as simple as selecting all the text in the PDF document, hitting "Edit/Copy", pasting it into a text file, and importing that text file into Excel (with the proper formatting instructions)... is it?