Beyond the CSV export
The export button is where good data goes to get stale. There's a better default than dumping rows into a spreadsheet.
Ryan Chandler
2 min read
There's a quiet ritual in most companies. Someone needs a number, so they export a CSV, open it in a spreadsheet, build a pivot table, and email a screenshot. A week later they do it all again, because the data has moved on and the spreadsheet hasn't.
The export button feels like freedom. Mostly it's a copy machine for stale data.
Why the spreadsheet wins anyway
We're not knocking spreadsheets. They win for a reason. They're flexible, they're familiar, and they let you ask the next question by typing into a cell instead of filing a request. That last part is the whole appeal.
The trouble is everything you give up to get there:
- The export is a snapshot, wrong the moment a new order lands
- The logic lives in formulas nobody else can see
- The numbers drift from the source as copies multiply
- "Final_v3_actually_final.xlsx" becomes the source of truth
A spreadsheet is a fork of your database that starts going stale the second you make it.
Keep the flexibility, lose the staleness
The thing people actually want from the export isn't the file. It's the ability to keep asking questions without waiting on anyone. Dronk gives you that against live data:
| CSV plus spreadsheet | Dronk |
|---|---|
| Snapshot, stale on arrival | Live query against the real database |
| Hidden formula logic | The exact SQL, visible every time |
| Re-export to refresh | Ask again, it's always current |
| Lives on one person's laptop | Pin it to a dashboard, share the link |
You still get a table you can read at a glance. You just get it from the source, with the query attached, and you can drill in with another plain-English question instead of another VLOOKUP.
Export is a fallback, not a workflow
There's nothing wrong with exporting a result when you genuinely need a file. Sometimes you do. The shift is making that the exception instead of the default path for every question.
Ask the database directly, get a live answer, keep going. When the data changes, your answer already has. That's the part the CSV could never do.