Skip to main content
BigQuery is the integration for teams whose real product data lives in their warehouse rather than in an analytics tool. It lets Superhawk compute metrics directly against your own tables, in your own definitions. This is usually the most accurate usage source available, because it measures the thing your business actually counts rather than an approximation assembled from client-side events.

​
Connecting

1

Create a service account in Google Cloud

Create a service account with read access to the datasets Superhawk should query, and the ability to run query jobs.
2

Generate a JSON key

Create a key for that service account and download the JSON.
3

Open Integrations in Superhawk

Go to Integrations and find BigQuery under Data Warehouse.
4

Paste the service account JSON

Paste the full JSON key. Superhawk verifies access immediately.
Grant the narrowest access that works: read on the specific datasets Superhawk needs, and nothing else. It never needs write access.

​
Defining metrics against it

Connecting BigQuery makes it available as a source. The useful work is defining the metrics that read from it. Go to Configure Metric and define each metric: what it measures, how it resolves to a single account, and what window it is computed over. See Metrics.
Always set the window explicitly, and preview before saving. A metric on the default window computes cleanly, returns plausible numbers, and measures a different period from the one you intended. Nothing errors, so a preview against an account you already understand is the only way to catch it.

​
Cost

Superhawk queries your warehouse on a schedule, and BigQuery charges by data scanned. Two things keep that reasonable:
  • Scope metrics to the columns and partitions they need. A metric that scans a whole table each run costs many times one that reads a partition.
  • Match the schedule to how fast the number moves. A metric that changes weekly does not need computing hourly.

​
What it feeds