Skip to content
How to ThinkIn the Age of AI

Verify data portability before committing

Before committing to a new tool, verify it stores data in exportable open formats by actually performing an export and checking whether another tool can import it without significant loss of structure or metadata.

When This Fires

Common Failure Mode

Feature-page trust: reading "Export to Markdown/CSV/JSON" on the tool's website and assuming it works well. The feature exists; the quality of the export is unknown until tested. A Markdown export that strips all YAML frontmatter, removes link targets, and converts tables into unintelligible plain text is technically "Markdown export" but practically useless.

The Protocol

(1) Create 10-20 representative items in the tool: notes with various formatting, items with metadata/tags, items with internal links, items with embedded content. (2) Export using every available export format. (3) Import the exported data into at least one other tool (ideally one you'd consider as a replacement). (4) Examine every imported item: Formatting — do headings, lists, bold, and tables survive? Metadata — are tags, dates, and properties present? Links — do internal references resolve? Content — is anything missing or garbled? (5) Grade: Clean export (all items import with minimal loss) → tool passes portability check. Lossy export (items import but with formatting/metadata loss) → acceptable if losses are fixable via script. Non-portable export (items are unusable after import) → tool fails portability check. Reconsider commitment per Choose open portable formats (Markdown, CSV, JSON) over….