Hello,
thank you for the report. With Force OCR, the online tool recognizes the pages again, but it adds the new text as an additional text layer. The text from your scanner stays in the file, so when you copy or search, you often still get the old text. The online tool currently has no option to remove the existing text first. We have passed this on to development.
Until then you can use PDF24 Creator on Windows: open PDF OCR in the PDF24 Toolbox, turn on Remove existing text and turn off Skip pages with text. The existing text is then removed before the recognition, so the result only contains the new text. Please note that this option removes all text of the PDF, so use it for scanned documents only, where the text is just the hidden text of the scanner.