Read only new parts of a file?

view story

http://unix.stackexchange.com – I have a large CSV file, that grows by some entries every day (created by an external software which writes always to this one file, no splitting possible). For performance reasons and further processing I only want to read the new data. Is there some command line tool, that allows me to define a starting offset for reading from a file and works good for these things? (HowTos)