Historical Import

A historical import is loading past events from a file or a data warehouse into the tool, keeping their original occurrence time rather than the load time.

A historical import usually happens when migrating from one tool to another or to fill a period before the tool was installed. The key to doing it right is separating event time from insert time: if last year's event enters with today's time, the whole historical trend breaks and everything piles onto one day, so the import must preserve each event's original time. Two more caveats: user identity must stay consistent between old and new data or one person is counted twice, and imported events often arrive out of order because they come all at once in bulk. The misconception is that a historical import fully reconstructs the past; only what was stored in the raw source comes back, and behavior that was never recorded returns with no import at all.