Data Minimization

Data minimization is a privacy principle stating that you should collect only the data actually necessary for your stated purpose, not everything you could possibly collect.

The logic of minimization is both ethical and practical: data you did not collect cannot leak, cannot be misused, and carries no retention or compliance cost. The common trap is the “might be useful later” mindset that gradually builds an unused store of sensitive data that is pure risk. But minimization is in tension with analyzability: collect too little and you later cannot ask a new question because you lack the data. The right balance is not the least data possible but the least data necessary for a clear purpose. For web analytics this means collecting at an aggregated level — country instead of a precise address, device type instead of a full fingerprint — wherever that suffices for the decision.