Your Notion CRM hums along smoothly until you hit around 500 rows. Then everything slows. Filters lag. Linked records flicker. Formulas timeout. And when you finally export to move to something faster, you open the CSV and realize: your relationship graph is gone, your formula fields are blank, and your lookup columns never made it across. Notion wasn't built as a CRM. It's a note-taking tool that learned to fake it. At scale, the gaps show. This playbook tells you exactly what exports clean, what orphans forever, and the 14-day audit-and-reconcile sequence that gets your data across intact. Why Notion hits the wall at 500 contacts Notion's performance degrades in three ways as your contact database grows: Database filtering slows exponentially. At 500+ rows, even simple filters ("status = active") introduce 2–3 second delays. Relational properties struggle. Lookups and rollups that fetch data from linked records start to stall, especially if your link field points to multiple contacts. Formulas and computed fields time out. If your "pipeline value" formula rolls up revenue across 10+ related deals, Notion stops calculating it reliably above 800–1,000 rows. This isn't a hard cap—Notion will let you store 10,000+ rows. But the usability cliff is around 500. Beyond that, your sales team spends more time waiting for screens to load than actually selling. The export reality: what orphans, what survives When you export a Notion database as CSV, here's what happens to your core CRM fields: Fields that export cleanly Text, email, phone — All intact. No issues. Single select and multi-select — Export as comma-separated text. Importable, but you lose the visual color tags. Checkbox — Exports as true/false. Readable. Date — Exports as ISO 8601 format (YYYY-MM-DD). Re-importable if the target platform respects the format. URLs and file attachments — URLs export as text links. Files require manual reattachment. Created/Last edited timestamps — Export correctly, useful for audit trails. Fields that export as blank or corrupted Relation/linked record properties — Export as blank. If you link a Contact to 5 Deals, the CSV shows nothing. You lose the graph entirely. Rollup fields — Blank. A "Total pipeline value" rollup that sums linked deal amounts? Gone on export. Lookup fields — Blank. A lookup that retrieves "deal owner name" from a related deal? Orphaned. Formula fields — Most formulas export their _result_ if they're simple (IF, ADD), but complex multi-step formulas often return errors or blanks. Count and summarize — Blank. Any aggregation computed across related records disappears. Files and attachments in relation — If a file is stored in a linked database, it's not accessible from the export. The core issue: Notion's export doesn't reconstruct relational depth. It flattens your database. Any value computed by relating two tables dies on export. The 14-day pre-migration audit Before you export anything, spend 3–4 days mapping your data dependencies. This audit prevents surprises during import. Days 1–2: Inventory your Notion structure List every database. Contacts, Deals, Companies, Tasks, Notes, custom tables—everything. Screenshot the field names and types. Identify every relation. Go to each "Relation" or "Linked record" field and note what it links to. Example: "Contact → Deals (many-to-many)." Flag every formula, rollup, and lookup. For each computed field, write down the logic. Example: "Total pipeline value" = SUM of deal amounts where deal owner = this contact. Count contacts per linked table. How many deals per contact (average and max)? How many companies per contact? This tells you the complexity of what you're losing. Days 3–4: Audit the data quality Run a "relation completeness" check. Filter Contacts to find how many have at least one linked deal. How many are orphaned (no deals, no companies)? These might not import at all if your new system requires relational context. Audit formula accuracy. Pick 5–10 contacts and manually verify their formula results. If "Total pipeline" shows ₹5L but you count only ₹3L in the deals, the formula is wrong and your data is already corrupted. Check for duplicate records. Search for contacts with the same email, phone, or company name. Use Notion's "Find duplicates" feature (available in paid plans). Duplicates will cause reimport errors. Validate custom fields. If you have single-select fields like "Status" or "Lead source," count the unique values. A "Status" field with 47 different values (instead of 5–6) signals inconsistent data entry. Days 5–7: Build your mapping document Create a spreadsheet with three columns: Notion field name — Exactly as it appears in Notion. Target field in [new CRM] — Where it will go (e.g., "Email → Email", "Status → Deal stage"). Recovery plan — For orphaned fields (relations, rollups, lookups), note how you'll rebuild them post-import. Example row: Notion: "Total pipeline value" (Rollup) Target: n/a — will recalculate in new CRM Recov