Data warehouses & BI (BigQuery, Snowflake, PowerBI, Sigma)

When a goal's value lives in your warehouse (ARR, MAU, churn, anything you'd usually report on from a dashboard), connect the warehouse and point a goal at the query or report. The displayed number stays in sync without anyone having to hand-update it.

Supported warehouses & BI tools #

Warehouses (run SQL directly)

BigQuery and Snowflake. Brev runs queries against your warehouse and stores only the result, not the underlying rows.

BI tools (pin a saved view)

PowerBI and Sigma. Brev pulls a value from a dataset or report you've already built in the BI tool, the same number you'd see if you opened it.

Looker, Metabase and Tableau are in the catalog but can't be connected yet. Their cards show Request access, so you can tell the Brev team you want them.

How to connect #

1. 1
   Find your warehouse in the Data section of Integrations

   Each tool has its own card. Click Connect on the one you use.
2. 2
   Provide credentials

   BigQuery and PowerBI use OAuth: Brev redirects you to sign in with Google or Microsoft. Snowflake offers two methods, picked with the Access Token and Key Pair tabs on the Connect dialog. Access Token: paste your Account URL (the https://….snowflakecomputing.com address) and a Programmatic Access Token (PAT) you generate in Snowsight. Key Pair: paste your Account URL, your Snowflake Username, and your Private Key (PKCS#8 PEM; Private Key Passphrase is optional, for an encrypted key), after registering the matching public key on that Snowflake user. Connect stays disabled until the required fields are filled. An address that isn't an https Snowflake URL, or a key Brev can't read, is turned down before Brev contacts Snowflake. Key pairs don't expire, so they're the better fit when your account disables PATs or you want a long-lived service-account connection. Sigma takes a client ID and secret from Sigma's Developer Access page plus your region (AWS US, GCP, or EU). Brev only talks to Sigma's own https://….sigmacomputing.com API addresses, so a Sigma connection saved with any other address stops refreshing and asks you to reconnect and choose your region. Brev never requires write access; if the credential's role happens to have it, Brev still doesn't use it.
3. 3
   Choose where the data lives, per goal

   There's no default to set on the Integrations page. When you connect a goal, you pick where its data lives: for BigQuery a project, dataset and table, for Snowflake the warehouse that runs your queries.
4. 4
   Connect a goal to a query or dashboard tile

   Open a goal that should be data-driven and click 'Connect to integration'. Paste the query (for warehouses) or pick a dashboard tile (for BI tools). Brev runs it once to confirm a number comes back, then schedules it.

Where the data appears #

• KPI values: connected KPIs show live values pulled from the warehouse. The KPI row on every goals view reflects the most recent query result.
• Goal trajectory: every scheduled refresh writes a data point onto the goal's chart. Over time the trajectory line is built entirely from warehouse data, not manual check-ins.
• AI Goal Insights: Insights commentary picks up on trends, dips, and acceleration in the connected metric.
• Suggested check-ins: when a metric crosses a threshold (you set), Brev pre-drafts a check-in for the goal owner.

How often it refreshes #

• Warehouse and BI tool connections refresh every 30 minutes from 11:00 to 23:59 UTC and every 4 hours overnight.
• Submitting a check-in on a connected goal triggers an on-demand refresh of that goal's query right then.
• Click the refresh icon on a goal's connection to re-run its query right away.

What this costs (in your warehouse) #

Each scheduled refresh is one query in your warehouse. For a goal connected to a simple aggregation, this is typically negligible. For a goal connected to an expensive view, multiply that cost by 29 a day (every 30 minutes for 13 hours, plus 3 overnight runs) × number of days you'll track it. If your warehouse is on per-query billing (BigQuery on-demand, Snowflake compute time), factor that in when picking what to anchor goals on. Consider materializing expensive queries into a daily-refreshed table and pointing Brev at the table instead.

Common issues #

Query failed

Open the goal's integration connection. Brev shows the exact error your warehouse returned. Most common: missing permission on a new dataset, query timeout (raise the timeout in your warehouse), or a column that was renamed.

Connect says the token is invalid, but you're sure it's right (Snowflake)

If your Snowflake account has a network policy, it blocks Brev's servers before the token is checked. Snowflake returns the same 401 it uses for a bad token (error code 390422). Add Brev's egress IP addresses (shown in the Connect dialog) to the network policy allowlist, then reconnect. The token itself doesn't need to change.

Result is a row instead of a number

Brev expects a single scalar value per query (one row, one column, one number). If your query returns multiple columns or rows, wrap it in SELECT … FROM (your-query) or pre-aggregate.

The KPI preview says the query doesn't return a number

The query ran, but its value is text (a name, a date, a label) rather than a number, so Connect stays disabled. Change the query to select a numeric column or aggregate, for example COUNT(*) or SUM(amount). The same check applies to a Salesforce field, a Sigma metric or a HubSpot value.

I'm seeing a number that disagrees with my dashboard

Most likely the dashboard and Brev's query are filtering differently (different date window, different segment). Open the dashboard's underlying query and the goal's connected query side by side. If they're identical and still disagree, click the refresh icon on the goal's connection.