What is e-commerce analytics, and how do you build it in Metabase?
E-commerce analytics turns the activity in your store — orders, products, customers, discounts, and refunds — into shared metrics about how much you sell, how profitably, and how often buyers return. In Metabase, you build it by getting your store data into a SQL database, modeling a small set of clean tables, and standing up dashboards anyone can read.
TL;DR — Almost every store shares the same shape: orders, line items, products, customers, discounts, and refunds. Model that shape once and most metrics and dashboards port across platforms. Metabase reads SQL databases — hosted platforms (Shopify, BigCommerce) sync in first, while WooCommerce already lives in MySQL and can be queried directly.
What does e-commerce analytics measure?
It measures demand, profitability, and retention — not vanity traffic. The durable, leader-friendly questions are:
- How much are we selling, and is it growing? (net revenue, orders, AOV)
- Are buyers coming back? (repeat-purchase rate, cohorts)
- What's actually profitable? (discount depth, refund rate, margin)
- What's selling and what's stuck? (product performance, inventory turnover)
- Where does the funnel leak? (conversion rate, cart abandonment)
Avoid vanity metrics (raw pageviews, gross revenue with no returns netted out). They're easy to inflate and rarely change a decision.
Which tools feed e-commerce analytics?
The same pattern applies to every store platform. Per-tool setup lives on each integration page:
| Tool | Best for | Getting data into Metabase |
|---|---|---|
| Shopify | Hosted stores of every size | Managed connector, dlt, or Admin API; MCP for exploration |
| WooCommerce | Self-hosted WordPress stores | Query the MySQL database directly (read replica) |
| BigCommerce | Open-SaaS and headless builds | Connector, dlt, or REST/GraphQL API; managed MCP |
| commercetools | Composable, API-first commerce | Export API, GraphQL, or dlt; official Commerce MCP |
| Medusa | Open-source headless commerce | Query the PostgreSQL database directly (read replica) |
| VTEX | Enterprise commerce & marketplace | OMS and Catalog APIs, or dlt; official Developer MCP |
| Amazon Seller | Amazon marketplace sellers | SP-API via connector or dlt; managed MCP |
| Etsy | Handmade & craft marketplace sellers | Open API v3 via connector or dlt; managed MCP |
| eBay | eBay marketplace sellers | Sell APIs via connector or dlt; managed MCP |
| Klaviyo | Email/SMS marketing attribution | API or managed connector; official MCP with OAuth |
| Gorgias | E-commerce support tied to orders | Sync alongside store data to link support and revenue |
What is the shared commerce data model?
Almost every store maps onto these entities. Model them as clean tables, not raw connector JSON:
| Concept | Common terms | Used for |
|---|---|---|
| Order | Order, sale, transaction | The unit of sale |
| Line item | Line item, order product, order item | Product and basket analysis |
| Product | Product, variant, SKU | Catalog performance |
| Customer | Customer, contact, buyer | Repeat purchase, cohorts |
| Discount | Discount, coupon, price rule | Promo depth and ROI |
| Refund | Refund, return, credit | Net revenue |
The single most important decision is a consistent definition of net revenue and a fixed set of order statuses that count as "paid." With them, every card agrees; without them, sales numbers drift between reports.
Which e-commerce metrics matter most?
Define each one once and reuse the definition everywhere:
- Net revenue — gross minus discounts and refunds; state your stance on taxes and shipping.
- Average order value (AOV) — net revenue ÷ orders, computed at the order grain.
- Repeat-purchase rate — share of customers with two or more paid orders in a window.
- Conversion rate — orders ÷ sessions; needs storefront/analytics data joined in.
- Revenue per visitor — net revenue ÷ sessions; ties demand to traffic.
- Cart abandonment rate — abandoned ÷ started checkouts; needs checkout data.
- Refund rate — refunded amount ÷ gross; break out by product.
- Inventory turnover — units sold ÷ average units on hand; needs inventory snapshots.
How do you connect a store to Metabase?
Three routes, depending on the platform:
- Direct database — WooCommerce lives in MySQL, so Metabase queries a read replica directly, no ETL.
- Pipeline route (warehouse-backed) — sync hosted platforms (Shopify, BigCommerce) into a database with a managed connector, dlt, or the API, then build durable dashboards.
- MCP route (AI-assisted) — pair the platform's MCP server with the Metabase MCP server for live, exploratory questions. Treat it as exploratory, and remember it creates no history.
Which dashboards should you build first?
- Store overview — net revenue, orders, AOV, and repeat-purchase rate in one exec roll-up.
- Product / catalog performance — top SKUs, units, and refund rate.
- Customers & cohorts — repeat purchase and revenue by acquisition month.
- Discounts & refunds — promo depth and where revenue leaks.
Common mistakes
Reporting off raw connector tables.→ Model a thin clean layer first, with consistent order statuses and one definition of net revenue.
Reporting gross revenue as net.→ Subtract discounts and refunds; be explicit about taxes and shipping.
Averaging line-item prices for AOV.→ Compute AOV at the order grain, or multi-item baskets skew the number.
Comparing week-over-week through seasonal peaks.→ Compare like periods — this Black Friday vs. last — instead.
History-dependent metrics without history.→ Inventory turnover and cart abandonment need snapshots and checkout events — caveat them otherwise.
Related
Integrations
FAQ
Does Metabase connect natively to Shopify or BigCommerce?
No. Metabase reads SQL databases and warehouses. Sync the platform's data into a database first, then connect Metabase to that database. WooCommerce is the exception — it already lives in MySQL, so Metabase can query it directly.
Can I use the same dashboards across store platforms?
Mostly yes, if you model each platform onto the shared commerce schema. Metric definitions and chart structure port across; only source-specific fields differ.
How do I measure conversion rate and cart abandonment?
Both need funnel data beyond orders. Join storefront sessions (from your web analytics) for conversion rate and revenue per visitor, and sync abandoned-checkout events for cart abandonment. Without them, caveat the metric.