Historical Data Revisions (Backfills)
What is a backfill?
Rockerbox has a limited ability revise historical data (also referred to as "backfilling" data) to modify either the structure (tiers) or the contents (ex attributed marketing events) of your data.
Limitations exist in order to:
1. Avoid erasing existing marketing touchpoints on a user's path to conversion that fall outside of the attribution window.
Ex: a user converted today and saw a Linear TV ad 12 days ago. If this data is revised in a week (19 days from the day of ad exposure) the Linear TV touchpoint will be dropped.
2. To avoid modifying performance data after weekly / monthly / quarterly reporting has been "locked in" and benchmarks have been set.
What are the limitations on backfills?
To avoid the above, Rockerbox limits revisions to historical data (or backfills) to:
- 30 days for any changes impacting attributed conversions or marketing events (ex ingesting a delay TV postlog from 2 months ago)
- 90 days for any changes impacting your spend data (or tier structure for customers leveraging the "Export" Data Warehouse product)
- Backfills to resolve minor discrepancies (make up < 1% of your total spend or marketing events) are not supported.