CSV vs JSON vs Excel: I've Wasted Hours Using the Wrong Format
I once spent 4 hours debugging a data import because I saved a file as CSV when the system expected JSON. The next week I sent a client an Excel file with formulas and they opened it in Google Sheets, which broke half the calculations. Data formats are boring until they waste your afternoon.
Use CSV when...
You need maximum compatibility. Every tool on earth reads CSV. Databases, spreadsheets, programming languages, even text editors. If you're exporting data that someone else needs to import, CSV is the safest bet.
Your data is flat. Rows and columns, nothing nested. Customer lists, transaction logs, survey results. CSV handles this perfectly. Use our CSV Viewer to browse large files without freezing Excel.
Watch out for: Commas in your data (use proper quoting or switch to TSV), Excel "helpfully" converting things (dates, long numbers, scientific notation), and encoding issues with non-English characters (always use UTF-8).
Use JSON when...
Your data has hierarchy. A user who has multiple addresses, each with multiple phone numbers? That's nested data. CSV can't represent this cleanly. JSON can.
You're working with APIs. Every modern API speaks JSON. If you're building integrations, thinking in JSON saves translation steps.
Watch out for: File size (JSON is 30-50% larger than equivalent CSV due to all the key names), readability (deeply nested JSON is a nightmare to debug by eye), and the fact that many non-technical stakeholders can't open JSON files without help.
Use Excel when...
Non-technical people need to read and edit the data. Your finance team isn't going to open a CSV in a code editor. Excel gives them filters, conditional formatting, pivot tables, and that comfortable spreadsheet interface.
You need formulas or formatting. CSV doesn't preserve formulas, colors, or cell formatting. If those matter, you need .xlsx.
Watch out for: Version compatibility (a .xlsx created in Excel 365 might not work perfectly in LibreOffice), file size (Excel files are larger than CSV for the same data), and the dreaded "someone edited the source Excel and broke the formula chain."
The decision in 5 seconds
Sending to a developer? JSON. Sending to a business person? Excel. Importing into a database or tool? CSV. Wrong choice? Use a converter and move on with your life.