How do you analyze creator platform data in Metabase?
Creator platforms each hold one slice of a creator business: Kajabi knows about offers, Thinkific knows about enrollments, Teachable knows about affiliate fees, Gumroad knows about digital sales, Patreon knows about pledges. None of them knows what you earned. A creator selling a course on Teachable, templates on Gumroad, and a membership on Patreon has three dashboards and no total — and none of the three subtract each other's fees, or join a sale to the email campaign and ad spend that produced it. In Metabase, land those records in one database and model orders, products, enrollments, members, and payouts once. That's where gross becomes net, and where five separate platform numbers become one P&L.
mb upload csv, and treat warehouse sync as the advanced tier.Which tools does this cover?
- Kajabi — transactions, orders and order items, offers and products, customers and contacts
- Thinkific — orders, enrollments and progress, users and students, products and bundles
- Teachable — transactions, enrollments and course progress, pricing plans, affiliates and authors
- Gumroad — sales, products and variants, subscribers and recurring charges, refunds, chargebacks, and disputes
- Patreon — members, campaigns and patron counts, pledge tiers and entitled amounts, charge dates and payment status
Payments and marketing sit next door and join on the same buyer email: Stripe for the processor's own view of fees and payouts, and Klaviyo for the email campaigns that produced the sales. For physical products alongside digital ones, see the e-commerce category.
What is the shared creator data model?
Five platforms, six tables. Every guide in this category maps its API responses onto the same names, so a card written against creator_orders works whether the rows came from Kajabi transactions, Teachable transactions, Thinkific orders, or Gumroad sales.
| Table | Grain | Used for |
|---|---|---|
creator_orders | One row per sale or transaction | Gross and net revenue, refunds, fees, average order value, repeat purchase |
creator_products | One row per course, offer, digital product, or pledge tier | Revenue by product, catalog mix, price and tier analysis |
creator_enrollments | One row per student per course | Completion rate, activation, cohort maturity, re-enrollment |
creator_lesson_progress | One row per student per lesson | Drop-off curves and where a course actually loses people |
creator_members | One row per member per daily snapshot | Active members, MRR, churn, cohort retention, failed payments |
creator_payouts | One row per payout to your bank account | Reconciling recognized revenue against money that arrived |
Two grain decisions do most of the work. creator_orders is per transaction, not per purchase — a three-payment plan writes three rows from one sale, so count distinct orders for volume and sum transactions for cash. And creator_members is a daily snapshot, not a member list: Patreon's and Gumroad's membership endpoints return current state only, so churn history exists only if you captured it, and it can't be recovered later.
How do you connect creator platforms to Metabase?
Metabase reads SQL databases and warehouses — it doesn't call SaaS APIs. So the job is getting platform records into a database Metabase can query, and for this category that means an export you schedule yourself in almost every case.
- CSV upload route (start here) — pull a slice from the platform's API with a short script, save CSV, and load it with
mb upload csv. This creates a table and a model in one command, andmb upload replacerefreshes it from a newer export. On Metabase Cloud there's nothing else to set up. Most creator businesses have thousands of orders, not millions, so this is a real answer rather than a demo. - Scheduled API pull (the normal route) — a nightly job against the platform API landing in Postgres: Kajabi's Public API for orders, transactions, and payouts; Teachable's Transactions API; Gumroad's
/sales; Patreon's Members endpoint appended with a snapshot date. Watch the per-platform limits — Teachable pages 20 records at a time under 100 requests per minute, Thinkific allows 120 per minute, Gumroad and Kajabi publish no limits at all, so page conservatively. - Managed connector (Thinkific only) — Fivetran's official Thinkific connector syncs orders, enrollments, users, products, coupons, and promotions, with enrollments arriving incrementally. Airbyte has a community source with 11 streams but no incremental sync, so schedule it accordingly.
- MCP for exploration, not for revenue — Thinkific's read-only beta server answers student and enrollment questions well; Kajabi's is genuinely useful for content and product operations. Neither will give you money data, by design.
What can you analyze across these tools?
- Platform fee rate— the effective take-rate, derived from fee lines rather than the rate card
- Average order value — by product and tier, with payment-plan installments handled deliberately
- Course completion rate — with a maturity window and free enrollments held out
- Repeat purchase rate — the earliest signal that a product ladder is working
- MRR — recurring pledge and membership revenue, normalized across monthly and annual cadence
- Churn rate — genuine cancellations, with failed payments tracked separately
- LTV — lifetime support per member, read as a median because a few whales distort the mean
- ARPU — average revenue per active member, which tier mix moves more than price changes do
Which dashboards should you build?
- Course revenue — gross, tax, fees, refunds, and net by product and month
- Membership growth — active members, net change, tier mix, and failed payments
- MRR / ARR — recurring revenue with monthly and annual pledges on one basis
- Churn & retention — cancellations and cohort retention from snapshot history
- Repeat purchase — second and third purchases across the catalog, keyed on buyer email
Common mistakes
Related
Analytics
Integrations
Dashboards
FAQ
What is creator platform analytics?
Do any of these platforms have a warehouse connector?
Can I use MCP servers to build creator dashboards?
Do I need a data warehouse to do this?
mb upload csv) turns a CSV export into a queryable table and a model in one command — mb upload replace refreshes it later. That covers a real gross-to-net dashboard across all five platforms. Move to a warehouse with scheduled syncs when you want history that survives without you, or when you're joining course sales to ads and email data. Every integration guide in this category documents both routes.Which tools does the creator platform category cover?
creator_orders, creator_products, creator_enrollments, creator_lesson_progress, creator_members, creator_payouts — which is what makes a combined revenue view possible at all.Why doesn't my platform dashboard match my bank account?
creator_payouts. See platform fee rate for how to derive the effective take-rate from fee lines rather than the published rate card.