Amazon S3 in Polar: drop a file, get a governed metric.

Drop any dataset into an S3 bucket and Polar turns it into governed metrics and dimensions, blended with Shopify, Meta and Google in one semantic layer. Offline and retail sales, 3PL and logistics costs, COGS tables, custom KPIs, any file your team already exports. No warehouse to run, no pipeline to build, no CSV to re-upload by hand.

/ ⏎

Connect Amazon S3 See the setup

Amazon S3 → Polar

Any file in a bucket, live next to Shopify and ads.

Point Polar at a bucket and drop your files in it. Their columns become queryable metrics and dimensions, blended with revenue, spend and orders, and refreshed on every scheduled sync. No manual re-upload, no data warehouse required.

Custom Datasets

“Can I import data no standard connector has?”

Drop a file in your S3 bucket and its columns become extra metrics and dimensions in Polar, available on the next scheduled data refresh. Whatever your team already exports, Polar can read.

file -> metrics & dimensions

Offline & Retail

“What about revenue that never touches Shopify?”

Wholesale, retail, marketplace and offline orders you export to a CSV land next to DTC, so the P&L reflects the whole business, not just the storefront.

offline orders + DTC, one P&L

Cost & Logistics

“Where do COGS and 3PL costs live?”

Drop cost and logistics files (COGS, landed cost, 3PL and shipping fees) into the bucket and blend them with Shopify revenue and ad spend for real contribution margin, not gross revenue.

COGS, 3PL & landed cost

Custom KPIs

“My KPI isn't in any connector.”

Whatever your team computes and exports (custom LTV, unit economics, proprietary metrics) becomes one governed Polar metric everyone reads the same way, no re-modeling.

custom KPI, governed once

Flexible Modeling

“How do my columns become usable?”

Polar maps your file's columns to custom dimensions and metrics, so a raw export turns into a governed dataset you can slice by SKU, channel, region or segment.

columns mapped to your schema

Blended Context

“How does file data sit with the rest?”

Your dropped files live in the same semantic layer as Shopify, Meta, Google and Klaviyo, blended into Net Sales, MER, CAC and margin, and readable by Ask Polar and the MCP.

files in the blended layer

Setup Guide

How to Connect Amazon S3 to Polar.

A Polar account with Shopify connected, an S3 bucket Polar can read, and the files you want to expose. Polar's team maps the columns into your semantic layer. Read access is enough, Polar never writes to your bucket.

01 · What you need

A Polar account with Shopify

Amazon S3 plugs in next to your store data, so the files you drop meet real revenue and spend automatically.

Connect your data →

An S3 bucket Polar can read

A bucket, and read access for Polar to it. Think of it as a shared folder: you drop files in, Polar reads them. Read-only is enough, Polar never writes back to your bucket.

Amazon S3 →

The files you want to expose

Pick the datasets to bring in: offline orders, a cost or 3PL file, a custom-KPI export.Amazon S3 sits under Custom connections available in Connectors. Click Request and Polar's AI Data Engineer builds it, in minutes.

02 · Connect it, the guided way

Request Amazon S3 in Connectors ~1 min

Custom connections available.

In Polar, open Connectors. Amazon S3 sits under Custom connections available: click Request and Polar's AI Data Engineer builds it, in minutes.

Open Polar connectors

Grant read access and drop your files ~5 min

A readable bucket and the datasets to expose.

Give Polar read access to your S3 bucket and drop the files you want in Polar. New files landing in the bucket sync on the next scheduled refresh.

Name the columns and custom KPIs to bring in. Polar maps each into your semantic layer as governed metrics and dimensions.

Best practice: start with one or two files you report on most (an offline-orders export, a cost file). Keep dropping new versions in the same bucket and Polar picks them up.

See it in your reports the payoff

File data, blended and governed.

Your dropped files now power dashboards, reports and every Ask Polar answer, blended with Shopify, Meta and Google. They feed contribution margin, true CAC and MER, with no pipeline to maintain and no CSV to re-upload by hand.

Side by side

Amazon S3 stores the data. Polar turns it into governed metrics.

Files land in Amazon S3 and stay files: no metric definitions, no blend with your store, and every question needs someone to write the query. Polar ingests them into the same semantic layer as your orders and your spend.

Amazon S3 on its ownon its own

Amazon S3 with Polarblended

Data sits as files and tables, with no metric layer on top.

Ingested into Polar's semantic layer as governed metrics.

Blending it with orders and spend means writing SQL.

Blended with Shopify, ads and email without writing SQL.

Every new question is a new query and a new export.

Ask in plain language, or read it in a dashboard that refreshes.

No shared definition, so two teams arrive with two numbers.

One definition, so the number is the same everywhere.

Featured Use Case

The whole-business P&L, online and offline in one file drop.

Your retail, wholesale and offline sales live in a spreadsheet, not Shopify. Drop them in S3 and blend them with DTC revenue and ad spend in Polar for a P&L that reflects the whole business. No warehouse, no reverse-ETL, no manual re-upload.

Revenue by channel type● offline from S3

Channel

Revenue

Margin

Shopify DTC

$182k

58%

31%

Retail (from S3)

$96k

30%

22%

Wholesale (from S3)

$38k

12%

16%

Offline and wholesale dropped in S3, blended with Shopify DTC in Polar.

Mapped once

No AI needed

Map your offline-sales file into the semantic layer once. Whole-business revenue and margin then show up in every dashboard, report and scheduled export automatically. Drop a fresh file and it just updates.

AI agent

Dynamic

Give an agent the Polar MCP. Ask it to split revenue and margin across DTC, retail and wholesale using your S3 files, and it queries the blended model in plain language, no SQL required.