How BigQuery and a customer data platform can work together
Connect selected BigQuery records with customer profiles and return configured audience data for reporting. Understand what each data flow needs to work.
Marketing may need an audience for the next customer journey, while finance and operations need a wider view of the business. Those needs can share information without putting every responsibility into the same platform.
A warehouse such as BigQuery can support analysis across business sources. PangoCDP brings customer information into profiles, audiences and configured journeys. The useful connection is the particular information one needs from the other.
Separate the two data flows
PangoCDP documents a BigQuery input connection and an optional BigQuery export within its advertising-audience synchronisation process. These are separate configurations with different purposes. A proposal should describe each direction, its fields and its timing.
| Direction | Documented purpose | What to establish |
|---|---|---|
| BigQuery → PangoCDP | Selected query results are mapped into an object model. | The query, record identifier, target fields and source update process. |
| PangoCDP → BigQuery | A configured audience synchronisation can also export the selected segment data to BigQuery. | The segment version, export configuration and reporting use. |
Bring a selected query into a customer model
The input guide describes selecting a BigQuery connection, writing a query for the required fields, checking its result and mapping the columns to a PangoCDP object model. The mapping includes a required reference identifier, shown as Ref Id, to distinguish records.
Decide what one returned row represents before building that mapping. A customer, an order and an order line are different records. Mapping every line of an order onto one customer identifier can produce repeated or conflicting updates.
Keep the first exchange focused. A journey may need a customer reference and a purchase date; it may have no need for supplier costs, financial notes or the entire order history.
Return an audience version for wider reporting
The documented outbound option sits within advertising-audience activation. The operator selects the segment version and schema, chooses the activation trigger and can enable the parallel BigQuery export. A scheduled configuration also distinguishes using the latest calculated version from recalculating before activation.
Define how the exported audience will be recognised in reporting. A campaign identifier, audience version and relevant timestamps help an analyst distinguish the selected population from later recipients, respondents and purchasers. Confirm which fields are actually available in the configured export.
With the necessary records and identifiers, the wider reporting layer can compare that population with operational outcomes. An audience export alone does not supply every delivery, response or transaction event, and it does not establish that advertising caused a later purchase.
Keep source responsibility and timing visible
The order system should remain authoritative for its transaction status. The customer programme needs an agreed source for preferences. The reporting team needs to know which query and audience version produced a result.
Review three timings separately: when the source changes, when the input data reaches the customer model and when the audience is calculated or exported. A fresh audience calculation can still use an old purchase record if the preceding exchange has not completed.
Name the people who maintain the query, approve its business meaning and investigate a failed or incomplete exchange. Include access, query cost, operating support and change testing in the implementation discussion.
Prove one useful exchange before expanding it
Trace a few representative records through the proposed design: an ordinary purchase, a cancellation, a changed customer identifier and a late source update. Check the mapped record and the resulting audience or exported version against the intended result.
Twin can shape the enterprise data foundation, the PangoCDP customer programme and the integration between them. The starting point is one useful question and a flow your teams can understand, operate and verify. For the wider division of responsibilities, see the CRM and customer-data comparison.