Windsor.ai vs a custom data warehouse: cheap connectors or a finished model?
Windsor.ai moves data from 350+ sources into sheets, Looker Studio, or Snowflake from $19 a month. It does not model profit. Here is how it fits with a build.
Linus TseGreyfield DataFacts last checked Contents 10 sections
The short answer
Windsor.ai is one of the cheapest ways to get marketing and store data into Looker Studio, a spreadsheet, or your own Snowflake, and every destination is on every plan. It is a pipe, not a model: it does not calculate profit, match orders across channels, or handle refunds for you. For many brands it can be the pipeline under a custom build.
How Windsor.ai and a custom warehouse differ
- 01
Connectors vs modeling
Windsor connects 350+ sources and delivers them as tables. Matching orders, dating refunds, and calculating margin happen after delivery.
- 02
Snowflake on every plan
Unlike several competitors, Windsor includes warehouse destinations on its cheapest plans, so it can feed a build at low cost.
- 03
Priced by sources and accounts
Plans run from $19 to $499 a month billed annually, driven by sources, accounts, and refresh frequency.
Pricing at a glance
Windsor.ai
unknown (independent company; ownership not verified)
Windsor.ai is a no-code data integration tool that pulls marketing, ecommerce and other SaaS data from 350+ sources into BI tools, spreadsheets, warehouses and AI assistants. It is aimed at marketers and small data teams who want connectors without running their own pipelines.
- Pricing model
- Tiered subscription by number of data sources and accounts
- Published starting price
- $19/month billed annually (Basic, $23 month-to-month)
- Free tier
- Yes: 30-day trial (10 sources, 15 accounts), then a Forever Free plan with 1 user, 1 data source, 1 account
- What drives cost
- Tiers (annual / monthly price per month): Basic $19/$23 (3 sources, 75 accounts, daily sync, 5 destination tasks); Standard $99/$118 (7 sources, 75 accounts, hourly, unlimited destination tasks); Plus $249/$299 (10 sources, 200 accounts); Professional $499/$598 (14 sources, 500 accounts, 15-minute refresh); Enterprise by quote (up to 200 sources and 50,000 accounts, SSO, white-label, custom connectors). Cost drivers: number of data sources, total accounts across sources, refresh frequency, destination tasks on the lowest tiers. Users unlimited on paid plans; all destinations on all plans. Changelog notes volume-based pricing for database destinations launched July 2025, but the pricing docs do not describe it. Setup time: vendor claims a connection in under 2 minutes; Shopify and Snowflake-side prep is extra and not timed.
Custom warehouse (Greyfield)
A Snowflake warehouse with dbt models and a BI layer (usually Sigma), built around your sources and your definitions, that you own after handover.
- Pricing model
- Project-based build, optional ongoing support
- Published starting price
- Scoped after a free consultation
- Free tier
- No
- What drives cost
- Warehouse, connector, and BI licences are billed by those vendors directly to you.
Feature comparison
"Not verified" means we could not confirm it from public docs, not that the feature is missing. Each cell links to the page we checked.
Sales channels
| Feature | Custom warehouse (Greyfield) | Windsor.ai |
|---|---|---|
| Shopify Native Shopify connector. | Yes Through a managed connector such as Fivetran or Airbyte. source | Yes Native connector; vendor cites 122 metrics and 313 dimensions source |
| Amazon Seller Central Native Amazon connector. | Yes Orders, settlements, and fee events. source | Yes Seller Central connector with orders, fees (referral, FBA, storage), inventory; Vendor Central also listed source |
| TikTok Shop Native TikTok Shop connector. | Yes Modeled alongside Shopify and Amazon. source | Yes Listed as a connector; field list not checked source |
| Wholesale, retail, ERP, 3PL Native connectors for sources beyond the online store, such as NetSuite, Faire, retail POS, or a 3PL. | Yes Any source with a connector or export. source | Yes NetSuite listed; Faire and ShipStation not listed source |
Marketing
| Feature | Custom warehouse (Greyfield) | Windsor.ai |
|---|---|---|
| Ad platform spend Meta, Google, and TikTok ad spend connectors. | Yes Ad spend modeled next to orders. source | Yes Facebook/Meta Ads, Google Ads, TikTok Ads and many others source |
| Tracking pixel + attribution Its own pixel and multi-touch attribution models. | No Attribution from a dedicated tool can be loaded in. source | Not verified Offers an attribution Looker Studio connector and dashboard template built on platform data; no own tracking pixel found source |
| MMM / incrementality Marketing mix modeling or incrementality testing. | No Not part of a standard build. | Not verified No MMM or incrementality product found on pricing, connector or attribution pages source |
Profit and customers
| Feature | Custom warehouse (Greyfield) | Windsor.ai |
|---|---|---|
| Profit with COGS and fees Contribution margin after product cost, shipping, and fees. | Yes Fees, refunds, and cost timing defined in dbt. source | Not verified Amazon fee fields exist, but no built-in profit or contribution margin model found source |
| LTV and cohorts Lifetime value, cohort, and retention analysis. | Yes Contribution-margin LTV by acquisition channel. source | Not verified Vendor mentions customer lifetime value data from Shopify; no cohort/retention feature verified source |
| Subscription analytics Recharge or other subscription reporting. | Yes Recharge renewals and skips modeled as events. source | Partly Recharge connector exists; no prebuilt subscription analytics verified source |
| Custom metric definitions Define your own metric logic. | Yes Definitions live in dbt models you own. source | Yes Formula-based custom fields, including cross-source calculations source |
Data access
| Feature | Custom warehouse (Greyfield) | Windsor.ai |
|---|---|---|
| Data in a warehouse you can query Your modeled data lands in Snowflake, BigQuery, or a hosted warehouse you can query with SQL. | Yes Your own Snowflake account. source | Yes Snowflake, BigQuery, Redshift, Databricks, Postgres and others on every plan source |
| SQL in the product Write SQL inside the product. | Yes Full SQL in Snowflake and in Sigma. source | Not verified Custom fields use formulas; no in-product SQL found source |
| AI assistant Natural-language questions about your data. | Partly Depends on the BI tool; we have built text-to-SQL access for a client. source | Partly Access via MCP server and connectors for Claude, ChatGPT, Copilot, Gemini; no in-app assistant verified source |
| Mobile app A native mobile app. | Partly Depends on the BI tool you choose. | Not verified Not found source |
What Windsor does well
Its connector list covers Shopify, Amazon Seller Central with fee fields, TikTok Shop, NetSuite, Recharge, and the main ad platforms. It has a free plan after a 30-day trial, unlimited users on paid plans, and a fast-moving AI story: an MCP server, connectors for Claude, ChatGPT, and Cursor, and, since May 2026, the ability to make changes in Facebook and Google Ads campaigns.
What still needs modeling
We found no built-in profit model, pixel, or MMM. Custom fields support formulas, but the rules that make ecommerce numbers agree still have to be written: counting each order once across channels, dating refunds, and replacing estimated Amazon fees with actual ones.
Where Windsor.ai is better
- Much cheaper to start than a custom build: paid plans begin at $19/month billed annually, and every plan includes all sources and destinations, including Snowflake and BigQuery
- Connector maintenance is outsourced: Shopify, Amazon Seller Central (including referral, FBA and storage fee fields), TikTok Shop, NetSuite, Recharge, Klaviyo and the main ad platforms are prebuilt, so nobody has to maintain API pipelines
- Vendor says a connection takes minutes; a custom Snowflake + dbt + Sigma project usually takes weeks before the first dashboard
- Ships ready-made report templates for Looker Studio, Power BI, Sheets and Tableau
- Strong AI access path: MCP server, Claude/ChatGPT/Copilot/Cursor connectors and, since May 2026, write actions on Facebook Ads and Google Ads campaigns
- Refresh down to every 15 minutes on Professional and Enterprise plans
Who should pick which
Pick Windsor.ai if
- You want marketing data in Looker Studio or Sheets for tens of dollars a month.
- You already run a warehouse and need cheap connectors into it.
- You want AI tools to read and act on your ad data.
Pick a custom warehouse if
- You need profit, refunds, and fees modeled across channels.
- You want tested business rules your team can read and change.
- You want reports built for finance and operations, not only marketing.
Recommendations by team
| If you are | Pick | Why |
|---|---|---|
| Small marketing team on Looker Studio | Windsor.ai | Low cost and no engineering needed. |
| Brand with a data engineer | Windsor.ai | Cheap connectors into Snowflake; your engineer models. |
| Brand that wants margin by channel | Both | Windsor loads Snowflake; the build models profit on top. |
Recent changes to Windsor.ai
- 2025-07 Launched Windsor.ai MCP server; volume-based pricing for database destinations introducedsource
- 2025-08 15- and 30-minute refresh for database destinations on Professional plansource
- 2025-10 Microsoft Fabric destination launchedsource
- 2026-01 Claude connector went live; Cursor MCP connector followedsource
- 2026-05 Claude Code plugin; MCP write actions for Facebook Ads and Google Ads campaignssource
- 2026-06 Per-connector MCP servers; Triple Whale connector addedsource
- 2026-10 eBay and Shopee Ads connectors added; Gemini Enterprise destinationsource
Questions people ask
Does a custom warehouse replace Windsor.ai?
Not necessarily. Windsor can be the connector layer that loads the warehouse. A build replaces the spreadsheets and dashboards that try to model profit on raw connector tables.
Can Windsor.ai send data to Snowflake?
Yes. Windsor lists Snowflake, BigQuery, Redshift, Databricks, and others as destinations on every plan.
Sources and corrections
We checked these pages on the dates shown. Products change quickly. If you work at Windsor.ai and something here is wrong or out of date, email linus@greyfielddata.com and we will fix it.
- Windsor.ai pricing and plans
- Windsor.ai pricing information (docs)
- Windsor.ai connectors
- Windsor.ai changelog
- Shopify to Snowflake integration
- Amazon Seller Central connector
- Custom fields documentation
- Data-driven attribution for affiliate ROI (blog)
- Ecommerce data warehouses
- Multi-brand ecommerce portfolio case study
- DTC apparel brand case study
Related comparisons
Funnel vs a custom data warehouse: marketing data hub or commerce model?
Funnel collects and harmonizes data from hundreds of marketing sources, with MMM and attribution as an add-on. Here is where it ends and a warehouse model begins.
Supermetrics vs a custom data warehouse: a data pipe or a finished model?
Supermetrics moves marketing and store data into Sheets, Looker Studio, or a warehouse. It does not model profit. Here is how it fits with a custom build.
Saras Analytics (Daton, Pulse, iQ) vs a custom data warehouse
Saras sells ecommerce ELT (Daton), prebuilt dashboards (Pulse), and an AI analyst (iQ). Here is which part replaces a custom build and which part feeds one.