Removing byte-level exact duplicate content reduced chunk volume by only approximately 3.3%, and did not reduce the number of reading sessions at all.