Skip to main content

​
How does Marrow refresh data?

Marrow keeps your connected store and ad data up to date with three kinds of sync:
  • Initial sync — After you connect an account, Marrow prepares the most recent window of data first (about the last 30 days) so you can start exploring quickly.
  • Historical sync — Marrow then backfills older data (typically up to about 180 days total coverage) in the background.
  • Incremental sync — Every day, Marrow pulls recent changes so your tables stay fresh for AI platforms, MCP, and Data Explorer.
All three use the same cleaning and write rules, so fields and keys stay consistent whether a row arrived from the first sync or from yesterday’s update.

​
What happens after I connect an account?

  1. You authorize a store or ad account in Integrations (read-only access).
  2. Marrow creates sync tasks for that connection and its scopes (marketplaces, stores, ad accounts, and so on).
  3. Initial sync runs first so recent data becomes available soon.
  4. Historical sync continues until the default history window is covered.
Historical sync tasks are expected to finish within 48 hours, after which you can query the prepared history normally. Large accounts or platform-side delays can take longer; check the connection’s sync status if something looks stuck.
Marrow prepares a standard history window when you connect. Arbitrary custom backfill ranges are not guaranteed for every table.

​
When does the daily incremental sync run?

Incremental sync is scheduled every day.
  • The system enqueues daily incremental work around 02:00 UTC.
  • For Amazon Seller and Vendor data tied to a marketplace, related pulls are aligned to about 2:00 in that marketplace’s local time zone. See Timezones for the marketplace map.
Most report-style tables refresh on this daily cadence. Exact readiness can lag the enqueue time because platforms must finish generating reports first.

​
Why does incremental sync look back several days?

Each incremental run typically re-reads about the last 7 days, not only “yesterday.” Platforms often revise or late-deliver metrics (orders, ads performance, settlements). Looking back a short window catches those updates. Marrow upserts by business keys, so overlapping refreshes update existing rows instead of creating duplicates.

​
How can I see whether my data is up to date?

In Marrow, open the connection or Data view for a Connector. You can typically check:
  • Sync status (for example syncing, ready, paused, or needs attention)
  • Data freshness / last successful sync
  • Which tables are ready and rough progress during the first sync
If authorization expires or a connection is paused, new incremental syncs stop until you reconnect or resume.

​
What should I do if my data looks late?

Common causes:
  • The source platform has not finished publishing that day’s report yet
  • The Connector authorization expired or was revoked
  • Sync was paused for that connection
  • The first historical sync is still running (especially right after connect)
Reconnect or resume the Connector if needed. If status looks healthy but data still does not match the platform console after a reasonable wait, contact Marrow support with the connection name and the table you expected to update.