Hi Julianna, good question!
You can choose the "
Experiment Duration" attribution model which will:
1) Ignore any metric conversion windows you may have set already, then
2) Aggregate all data for each user from their first exposure until the analysis run time (or experiment end date, whichever is earliest).