How to create the BigQuery tables and transfers with Detrics presets

Create the table groups with presets

For every data source, Detrics ships a set of presets: pre-built tables with what’s commonly used on each platform. For this demo’s dashboard:

• Google Ads: the campaign and budget preset. It has every field needed.
• Meta: the campaign preset.
• Pinterest: the campaign preset.
• GA4: the traffic acquisition preset (it comes with several tables; the dashboard uses one).

Clicking any table in the group previews how the data will look in the final BigQuery table, and you have full freedom to pick the fields you want to bring. There’s a lot more data available than what this demo’s schema uses.

Create the transfers

With the table groups ready, the transfers come next: click create transfer, pick the accounts (three, as a test, or sync all to sync every current and future account, something our customers use a lot), the destination the data goes to, and the update frequency: daily is perfect.

The same path repeats for any table group. There are more configuration options for the more technical folks (each table’s sync type, the incremental deduplication refresh window), but the presets already come configured right.

The historical load and the steady state

The first load can take a while: in the video, one transfer pulled half a million rows in 39 minutes, and with sync all it can stretch into days. This feature can download up to the last 6 years of data, which matters if you care about the historical record of the accounts you manage. When it finishes, you get an email.

The following downloads are much smaller: thanks to incremental deduplication, only the configured window gets refreshed (here, the last 12 days), which is the data still shifting day to day. It’s the most efficient approach: it minimizes BigQuery cost and latency, and enables real-time dashboards.

From each table’s name in Detrics, one click takes you to BigQuery: there sit the four transfers’ tables with all the data loaded, a classic paid media dataset ready for Claude to work with.