Why your CSV shows é and ’ in Excel, and how to fix it

Every character in a CSV is stored as one or more bytes, and the encoding is the rulebook that maps bytes to characters. UTF-8, which almost every modern system exports, stores plain English letters in one byte and accented letters, curly quotes and currency symbols in two or three. The CSV format has nowhere to declare which rulebook was used, so the program opening the file has to guess.

When you double-click a CSV, Excel on Windows usually guesses your system’s legacy code page, typically Windows-1252 on English and Western European setups. That code page reads every byte as a character of its own. So the two bytes of é are shown as two characters, à and ©, and the three bytes of a curly apostrophe become â, € and ™. The pattern is distinctive: à or  followed by an odd symbol almost always stands for an accented letter, and †followed by another character almost always stands for punctuation such as ’, “, – or •.

Two different situations look identical on screen. In the first, the file is fine and only the way Excel opened it is wrong, so reopening it correctly fixes everything. In the second, someone already opened the file the wrong way and saved it, or a system upstream did, and the garbled characters are now the real content of the cells. Reimporting can’t help then, because the file genuinely contains “Café”. To tell them apart, open the CSV in a plain text editor such as Notepad or TextEdit: if you see “Café” there, it’s the first case; if you see “Café”, it’s the second.

If garbled text goes through the same mistake a second time, each garbled character is garbled again, and é turns into a longer run such as é. The cause and the cure are the same, just applied twice.

SpreadsheetFix handles both situations. When you load a CSV, it reads the raw bytes and decides the encoding itself: a byte order mark is honoured, valid UTF-8 is read as UTF-8, and anything that isn’t valid UTF-8 is read as Windows-1252, the code page older Windows software writes. A file that was only opened the wrong way comes through clean, with no setting to choose.

For text that is already garbled, the Fix Encoding rule reverses the mistake instead of looking characters up in a table. It turns each character of a suspect cell back into the Windows-1252 byte it came from, then reads those bytes as UTF-8. If that gives valid text, the cell is replaced; if any character couldn’t have come from this mistake, the cell is left alone. Text that went through the mistake twice is repaired in a second pass.

The CSV you download is UTF-8 with a byte order mark, so double-clicking it in Excel shows the accents correctly. Every changed cell is listed in the preview before you export, and amount columns are covered by the totals check, so you can confirm that repairing payee names didn’t touch a single number.