Exports often store numbers as text: "$1,204.50", "12%", "15 kg". as.numeric() gives up on anything that isn't a plain number and returns NA, with a warning that is easy to miss.
parse_number() from readr reads the first number it finds and drops the currency sign, the thousands commas and any trailing unit: parse_number("$1,204.50") is 1204.5. For a percent it returns 12 from "12%", so divide by 100 if you need a proportion.
Check the result. A text column that turns into a number column should not gain missing values, so compare sum(is.na()) before and after, and look at any row that changed from a value to NA.
parse_number() reads the number out of the text; as.numeric() gives NA.