Free forever·No credit card·2-min setup·EU-hostedStart free
Upgrade · Spreadsheets

After the spreadsheet.

Excel and Google Sheets are brilliant. Most of this site was probably planned in one. The question is when your workbook has quietly become a database, a reporting engine, and a data warehouse, held together by VLOOKUPs and a prayer. This page is for that moment.

Signs you have outgrown it

The workbook has crossed 100k rows.
Sheets slows, Excel starts freezing on recalculation, VLOOKUPs take seconds.
Two people overwrite each other's edits.
Version history helps until someone deletes a tab and nobody notices for a week.
You have a tab named `DO_NOT_EDIT`.
The spreadsheet is now critical infrastructure with no access controls.
You rebuild the same pivot every Monday.
The reporting step is manual and error-prone. It should be a saved query.
Data lives in three tabs and a Google Form.
The glue between sources is copy-paste. One skipped Monday and the chart is wrong.
Someone asked for an API.
Spreadsheets don't expose clean REST endpoints. You can fake it; it will hurt.
The dashboard is `View-only` emailed screenshots.
The chart exists. The link you can share does not.

What you get on the other side

SQL, not nested IFs

Analytical questions in a readable language. Joins across sources. Aggregations that run in milliseconds, not minutes.

A real database under the hood

Your data sits in Apache Iceberg on S3, queried by DuckDB. Handles tens of millions of rows without thinking about it.

Shareable URLs

A chart is a link. Publish once, the number updates automatically. Email a URL; everyone sees the same thing.

An API for your ops team

Save a query, call it via REST. Scripts, dashboards, and internal tools pull fresh numbers on demand.

Version control for data logic

Transforms are named objects with a history. Change the SQL, the chart updates everywhere that query is used.

How to move your data over

Path 1: export a CSV, push it once

For a one-shot migration, export the tab as CSV and push it to a table. The schema is inferred. Querying starts immediately.

bash
# Install the CLI
curl -sSL rawquery.dev/install.sh | sh


# Push the CSV
rq push sales_q1.csv --table sales


# Query it
rq query "SELECT region, sum(revenue) FROM sales GROUP BY 1 ORDER BY 2 DESC"

Path 2: live Google Sheets connector

If the sheet keeps being updated (form responses, ops tracking, manual data entry), connect it directly. Every sync refreshes the table. Your spreadsheet stays the input surface; rawquery handles the queries and the charts.

bash
rq connections create ops-tracker \
  --type google-sheets \
  -p service_account_json=@sa.json \
  -p spreadsheet_id=1AbCdEfGhIjKlMnOpQrStUvWxYz
rq connections sync ops-tracker


rq query "SELECT status, count(*) FROM ops_tracker.tickets GROUP BY 1"

Column names are normalized (lowercased, snake_cased). Run rq describe ops_tracker.tickets after the first sync to see the exact names.

When staying in Sheets is the right call

  • Under ~10k rows. Sheets handles this comfortably. The overhead of a new tool is bigger than the problem.
  • The workbook is the workflow. Forms, manual edits, conditional formatting, Apps Script macros: this is what Sheets is for. Keep it.
  • Team is non-technical. If nobody on the team writes SQL and there is no intent to learn, Sheets with a well-structured summary tab beats a tool nobody touches.
  • It is working fine. A spreadsheet that ships correct numbers on time is a good spreadsheet. The time to move is when it stops doing that, consistently.

Try the new shape

Free tier includes enough rows and queries to migrate a real workbook and see if the shape fits. No credit card.

Start freeGoogle Sheets docsSee other comparisons