Tried to wget a webpage and hand it off to grep to get a string, but had trouble doing it because the file was produced by MS Office and was a UTF-16 file (Little-endian, windows line-endings, etc.) Not sure which part of all of that was the culprit, but a utility (looks like part of a stock Ubuntu install) called iconv will take care of this. Running this: iconv -f utf-16 -t utf-8 fixed.html |grep 'Symbola' worked a treat.