How does data retention differ from user retention?

Data retention is a storage setting expressed in time: until when the data stays available. User retention is a behavioural metric expressed as a percentage: what share of a cohort came back. The shared name is an accident of vocabulary and the two are computationally unrelated.

What is the right retention period?

There is no single number, and a figure quoted in another tool's documentation does not transfer to your situation. Three variables decide it: the longest analysis cycle actually run, such as a year-over-year comparison or a seasonal cohort; the legal obligations of the market you operate in; and the cost of storing raw versus aggregated data. The practical move is to set two separate ceilings, one for raw events and one for aggregates.

Data Retention

Intermediate

Updated Jul 11, 2026 Shahin editorial 2 min read

PronunciationEN/ˈdeɪtə rɪˈtɛnʃən/

Data retention is how long collected data stays available before automatic deletion, and the policy that sets that period for each class of data.

A retention period is normally defined per class of data: raw events, aggregate tables, session recordings and server logs each get their own ceiling, and aggregates usually outlive raw rows. The term should not be confused with user retention, which is the share of people who come back and has nothing to do with storage duration. Choosing the period is a trade-off: a longer window enables year-over-year comparison and long cohorts, a shorter one lowers legal exposure and storage cost. The worst case is a team discovering the ceiling on the day it needs a thirteen-month comparison and the raw rows are already gone.

FAQ