Shopify can't export to BigQuery on its own. Use a managed connector like Fivetran or Airbyte, a reporting tool that writes to BigQuery, or a custom pipeline on the Shopify GraphQL Admin API run on a schedule. Whichever you pick, load all order history with read_all_orders, sync by updated date, convert to the store timezone and model sales with Shopify's own definitions so totals match.
Shopify's own reports are fine until you need to combine them with something else: ad spend from Meta and Google, attribution data, Klaviyo, subscriptions, a second store. At that point the answer most growing brands reach is a data warehouse, and for teams already using Google tools that usually means BigQuery.
This guide covers the three ways to get Shopify data into BigQuery, how to choose, and the details that make the numbers match Shopify once they're there. It's based on pipelines we've built, including one that loaded more than 900,000 orders across two stores, described in our pet brand case study.
Why put Shopify data in BigQuery
- Join it with ad data. Calculate MER, new customer MER and contribution after ad spend from Shopify revenue, not platform claims. See how to find your true ROAS.
- Combine stores. Brands with a second store, a wholesale store or regional stores get one set of numbers.
- Keep full history. Your data stays queryable however your apps and plans change.
- Answer questions Shopify can't. Cohort retention by first product, repeat rate by acquisition channel, margin by SKU after discounts.
- Power dashboards and AI. Looker Studio, a custom dashboard or an AI assistant can query one clean source instead of five APIs.
The three ways to do it
1. A managed connector
Tools like Fivetran and Airbyte have Shopify connectors that copy your store's tables into BigQuery on a schedule. You authorize the store, pick a dataset, and the connector handles pagination, rate limits and schema changes.
Good for: getting raw data in quickly, especially alongside many other sources. Watch out for: pricing based on data volume that grows with your store, and raw tables that still need modelling before anyone can use them.
2. A reporting tool that writes to BigQuery
Some marketing data tools, Supermetrics for example, can send Shopify data to BigQuery as well as to Sheets or Looker Studio. They tend to deliver pre-shaped report tables rather than every field.
Good for: marketing teams who want reporting tables without engineering. Watch out for: less detail than the raw data, so order-level reconciliation can be hard.
3. A custom pipeline
A small program that calls the Shopify GraphQL Admin API, writes to BigQuery and runs on a schedule, for example as a Cloud Run job triggered by Cloud Scheduler. A bulk operation exports the full history once. After that, each run fetches orders updated since the last run.
Good for: full control over what's loaded and how, no per-row fees, and pipelines that also pull ad and attribution data the same way. Watch out for: someone has to build and maintain it.
Which option fits your store
| Managed connector | Reporting tool | Custom pipeline | |
|---|---|---|---|
| Setup time | Hours | Hours | Days |
| Running cost | Grows with volume | Monthly plan | Usually near zero on Google Cloud |
| Data detail | Raw tables | Report tables | Whatever you choose |
| Needs modelling | Yes | Less | Yes, usually built in |
| Maintenance | Vendor | Vendor | You or your partner |
| Best when | Many sources, an in-house analyst | Marketing reporting only | Specific needs, cost control, reconciliation |
Whichever you choose, the loading is the easy part. Most of the work, and most of the value, is in the next two sections.
Six things that break Shopify data in BigQuery
- Only 60 days of orders. Without the
read_all_ordersscope, the API only returns the last 60 days. Your history looks complete until someone checks last year. - Timezones. Shopify reports in the store's timezone. Load timestamps in UTC and convert to the store timezone before grouping by day, or every daily total will be slightly off.
- Orders change after they're created. Refunds, edits, cancellations and fulfillment all update an order later. Sync by updated date, not created date, and replace changed orders.
- Test and cancelled orders. Decide whether to exclude them and do it consistently, or totals won't match the Shopify reports your team already reads.
- Currencies. Orders have a shop currency and a presentment currency, the one the customer saw. Report in shop currency unless you deliberately need the other.
- More than one store. Each store has its own IDs, and some stores report the same store name. Label every row with the store's myshopify domain.
We build Shopify, ad and attribution pipelines into BigQuery in your own Google Cloud project, reconciled to Shopify before handover. Fixed quote, you own everything. Start with a free tracking check, or ask about a custom data build.
Get a free tracking checkFrom raw orders to numbers that match Shopify
Raw tables are not a report. Before anyone uses the data, build a daily sales model that follows Shopify's own definitions:
| Metric | Definition |
|---|---|
| Gross sales | Product price times quantity, before discounts and returns |
| Net sales | Gross sales minus discounts minus returns |
| Total sales | Net sales plus shipping, taxes and other charges |
| Returns | Counted on the day of the refund, not the day of the order |
Then check it. Pick a settled month and compare your model's totals with Shopify's sales report for the same dates. When they match, you have something the whole team can trust. When they don't, the gap is almost always one of the six problems above.
With Shopify reconciled, add ad spend and attribution data on the same daily grain. That's when the warehouse starts paying for itself: one table with spend, attributed revenue and real Shopify revenue for every day, which is how we found Meta claiming more purchases than a store took on 7 days out of 22.
What it costs to run
BigQuery itself is rarely the expense. Its free tier includes 10 GB of storage and 1 TB of queries every month. Hundreds of thousands of orders, plus daily ad data, typically stay well inside that, and a scheduled Cloud Run job that refreshes every couple of hours usually fits inside Cloud Run's free tier too.
What you pay for is the connector, if you use one, and the time to build the models and keep them correct. That's the trade to make: a managed connector costs money every month, while a custom pipeline costs once to build. Our BigQuery data warehouse service is a fixed-price version of the second, with dashboards on top when you need them.
Common questions
Can Shopify export data to BigQuery directly?
Not on its own. You need a connector such as Fivetran or Airbyte, a reporting tool that writes to BigQuery, or a custom pipeline that reads the Shopify Admin API and loads BigQuery on a schedule.
Why does the Shopify API only return orders from the last 60 days?
By default, apps can only read the last 60 days of orders. To load older history, the app needs the read_all_orders scope, which Shopify grants on request for apps that need it.
Should I use the REST or GraphQL Admin API?
GraphQL. Shopify treats the REST Admin API as legacy, and the GraphQL bulk operations are the fastest way to export a full order history in one job.
Why don't my BigQuery totals match Shopify's sales report?
Usually timezone, refunds or definitions. Shopify reports in the store timezone, records refunds on the day they happen, and separates gross sales, discounts, returns, taxes and shipping. Your model has to do the same to match.
How much does BigQuery cost for a Shopify store?
For most stores, very little. BigQuery's free tier includes 10 GB of storage and 1 TB of queries a month, and even hundreds of thousands of orders take up well under that. Connector fees usually cost more than BigQuery itself.
