Integration · BigQuery
Exporting GA4 to BigQuery moves the same consent-shaped data into a warehouse, with Consent Mode's modelled users left behind. The Sealmetrics connector writes traffic, conversions and revenue by channel, campaign and creative into your own BigQuery project, hourly or daily, as ready-to-query tables in an EU or US dataset you choose.
Every plan, free Agentic tier included · star schema · hourly or daily sync · your data, your retention
Quick answer
The Sealmetrics BigQuery connector exports your analytics data into a BigQuery dataset in your own Google Cloud project. You create a service account with the BigQuery Data Editor and Job User roles, upload its JSON key in the site's integration settings, choose the dataset name and an EU or US location, and pick a sync frequency: hourly, daily or manual. The data arrives as a star schema of fact tables — daily and hourly traffic, conversions, microconversions, pages and landing pages — partitioned by date and carrying UTM source, medium, campaign, term and content, channel group, country and device. Conversions include revenue and their custom properties as JSON. You can backfill history on the first run and later. The tables are aggregates without user identifiers, so they answer channel and revenue questions, not user-level ones. The connector is included on every plan, the free Agentic tier among them; Google bills storage and queries.
What it exports
You choose which data types to sync. Each one maps to a dedicated table with a fixed star-schema name, and every row carries the site's account ID.
| Data type | Table | What it holds | Note |
|---|---|---|---|
| Daily traffic | fact_traffic_daily | Entrances, engaged entrances, pageviews, microconversions, conversions and revenue by UTM, channel, country, device, browser and OS | Recommended; partitioned by date, clustered by account, source and country |
| Hourly traffic | fact_traffic_hourly | The same traffic metrics, hour by hour | Optional; 90-day partition expiration and higher storage cost |
| Conversions | fact_conversions | Conversion type, count, amount, revenue, attributed UTM and channel, landing page, click ID and custom properties as JSON | Recommended; clustered by account, conversion type and source |
| Microconversions | fact_microconversions | Microconversion events such as add to cart or sign-up | Recommended |
| Pages | fact_pages | Pageviews and entrances by page and content group | Recommended |
| Landing pages | fact_landing_pages | Performance by landing page | Recommended |
| Account metadata | dim_accounts | Site details | Optional; dim_countries and sync_metadata are always created |
Attribution in the tables is the same last-click attribution as in the dashboard: each conversion carries the source of the session in which it happened. The complete schema for every table is shown on the integration screen. How that attribution works is covered in campaign revenue attribution.
What a GA4 export does not fix
Teams often move to BigQuery to escape GA4's interface limits. The export solves that, and inherits everything upstream.
The GA4 export contains what GA4 collected. Modelled users and sessions from Consent Mode are not available in data exports, so a warehouse sees the consented share only. On Incapto's Shopify store, GA4 did not record 29% of visits. Why modelling does not close it is explained in Consent Mode: measured vs modelled.
Standard GA4 properties have a daily BigQuery export limit of 1 million events. High-traffic stores reach it, and the rest of the day is not exported.
The GA4 export is event-level, so channel, session and revenue reports have to be reassembled in SQL before anyone can use them. The Sealmetrics tables arrive already aggregated by channel and campaign.
Set up the connector
Part of the work is in Google Cloud and part in Sealmetrics. You own the project, the dataset and the retention from the first sync.
In Google Cloud Console, open IAM & Admin → Service Accounts, create an account such as sealmetrics-export, then add a key in JSON format and save the downloaded file securely.
Give the service account the BigQuery Data Editor and BigQuery Job User roles on the project, or a custom role that can create datasets and tables, update table data and create jobs.
Open the site in Sealmetrics, go to Site Config → Integrations → BigQuery and upload the JSON key. Sealmetrics checks it is a service account key and fills in the Google Cloud project ID.
Set the dataset name and its location, EU or US, pick hourly, daily or manual sync, select the data types to export and choose how much history to load on the first run.
Save the configuration, click Setup Dataset if prompted, then Sync Now. The integration screen shows the status, the last sync, the tables created and a history of rows and bytes per run.
Query it
Taken from the connector documentation. Replace your-project and sealmetrics with your project ID and dataset name, and filter on date so each query scans only the partitions it needs.
Entrances, pageviews and conversions per day for the last 30 days.
SELECT
date,
SUM(entrances) AS entrances,
SUM(page_views) AS page_views,
SUM(conversions) AS conversions
FROM `your-project.sealmetrics.fact_traffic_daily`
WHERE date >= DATE_SUB(CURRENT_DATE(), INTERVAL 30 DAY)
GROUP BY date
ORDER BY date DESCConversions, revenue and average order value by source and medium for the last 30 days.
SELECT
utm_source,
utm_medium,
SUM(count) AS conversions,
SUM(revenue) AS revenue,
SAFE_DIVIDE(SUM(revenue), SUM(count)) AS avg_order_value
FROM `your-project.sealmetrics.fact_conversions`
WHERE date >= DATE_SUB(CURRENT_DATE(), INTERVAL 30 DAY)
GROUP BY utm_source, utm_medium
ORDER BY revenue DESCPageviews and entrances by content group for the last 7 days.
SELECT
content_grouping,
SUM(page_views) AS page_views,
SUM(entrances) AS entrances
FROM `your-project.sealmetrics.fact_pages`
WHERE date >= DATE_SUB(CURRENT_DATE(), INTERVAL 7 DAY)
AND content_grouping IS NOT NULL
GROUP BY content_grouping
ORDER BY page_views DESCConversions and revenue broken down by a property sent with each conversion, here customer_type.
SELECT
JSON_VALUE(properties, '$.customer_type') AS customer_type,
SUM(count) AS conversions,
SUM(revenue) AS revenue
FROM `your-project.sealmetrics.fact_conversions`
WHERE date >= DATE_SUB(CURRENT_DATE(), INTERVAL 30 DAY)
GROUP BY customer_type
ORDER BY conversions DESCWho uses the export
The connector is for teams whose analysis already lives in the warehouse, next to spend, orders and finance data.
Join analytics with the rest of the warehouse.
Revenue by channel and campaign next to ad spend, CRM stages or product margins, without an ETL job to maintain.
How to measure ROAS after consentChannel revenue beside the ERP.
Conversions and revenue per day in a table finance can reconcile against booked orders.
One number for marketing and financeAggregate inputs for models.
Daily outcomes by channel and campaign are the input a marketing mix model or an experiment readout needs, with no user data.
Last-click vs modelled attributionDashboards in the tool the client uses.
Data Studio, Power BI, Tableau or any BigQuery-compatible tool reads the tables directly.
Campaign revenue attributionWhat the tables carry
No customer case about the BigQuery connector is published. What the export carries is the same measurement the published cases rely on: on Incapto's store, that base recorded 96% of real Shopify orders before anything else was compared.
of real orders recorded in Incapto's 48-day parallel run, the base the export is built on
data types you can sync, each to its own table, plus two tables always created
events a day: the BigQuery export limit on a standard GA4 property
What it does not do
These follow from what Sealmetrics collects. What is and is not stored is set out in demonstrate compliance.
The tables are aggregates without user identifiers. User-level paths, cohorts and retention cannot be built from them.
Conversions carry the source of their own session. There is no multi-touch or data-driven attribution to export.
The hourly traffic table has a 90-day partition expiration. Use the daily tables for longer history.
Storage and queries are charged by Google Cloud to your project. Filter on date and sync only the data types you need.
The dataset can be created in the EU or the US. Pick the EU location if the data has to stay in the EU.
Sync runs hourly at most, and recent data can take one to two hours to be processed before it appears.
Common BigQuery questions
Create a Google Cloud service account with the BigQuery Data Editor and BigQuery Job User roles and download a JSON key. In Sealmetrics, open Site Config → Integrations → BigQuery, upload the key, choose the dataset name, location, sync frequency and data types, then set up the dataset and run Sync Now.
The connector is included on every Sealmetrics plan, from the free Agentic tier to Growth, Scale and Enterprise, at no extra charge from Sealmetrics. Google Cloud bills your project directly for BigQuery storage and queries.
One table per data type you select: fact_traffic_daily, fact_traffic_hourly, fact_conversions, fact_microconversions, fact_pages, fact_landing_pages and dim_accounts. A dim_countries lookup table and a sync_metadata table are always created. Tables use a star schema, partitioned by date.
No. The tables are aggregates by date, UTM source, medium, campaign, term and content, channel group, country, device, browser and OS, with no user identifiers. Conversions include their count, amount, revenue, landing page, click ID and custom properties as JSON.
You choose hourly, daily or manual sync. On the first run you can backfill the last 30, 60 or 90 days, a custom number of days or a date range, and you can run larger historical loads later from the backfill card.
Yes, if you choose the EU location when you configure the dataset; the US location is also available. Sealmetrics' own analytics data is hosted in Dublin, and once data is synced into your project, you control its retention.
The GA4 export is event-level and contains what GA4 collected: consent-gated data, without Consent Mode's modelled users, and on standard properties up to 1 million events a day. The Sealmetrics export contains aggregates measured without consent loss, already organised by channel and campaign, with no user-level rows.
Deleting the integration removes the configuration from Sealmetrics but never deletes data already in your BigQuery project. Sealmetrics only pushes data; the one table with automatic expiry is fact_traffic_hourly, at 90 days.
BigQuery walkthrough
Book 30 minutes with the founder. We connect Sealmetrics to your BigQuery project, run the first sync and write the first query against your own spend or order tables.