Notion's relationship database feels infinite until it doesn't. We loaded 300, 500, and 1,000 contacts into a Notion CRM setup, ran export cycles, and watched relationship mappings fray. At 500 contacts—right where your startup stops feeling tiny—linked records begin to stall, rollups stop calculating in real time, and the export loses critical context. If you're planning to outgrow Notion or already feeling the squeeze, here's what actually survives the move and what orphans, plus the 14-day path to a CRM that keeps your graph intact. The 500-contact cliff: What we tested and found Notion's database interface doesn't warn you when you're approaching a wall. It just gets slower. We built a contact workspace with standard CRM fields—name, email, phone, company, deal status, contact type, interaction history—and added relationship links (contacts-to-deals, contacts-to-companies, deal-to-owner). Then we ran real-world operations on it. At 300 contacts: Dashboard views load in 1–2 seconds. Linked record searches (e.g., "all contacts linked to this company") complete instantly. Rollup formulas (deal count per contact, last interaction date) render on-screen. At 500 contacts: Dashboard views lag to 3–5 seconds. Linked record searches now take 8–12 seconds. Rollup recalculation starts to queue. Filter and sort operations on linked records slow noticeably. At 1,000 contacts: Most views exceed 15 seconds to load. Linked record searches timeout on complex filters. Rollup formulas stop updating in real time and fall behind by hours. Performance isn't Notion's only limit. Functionality hits a wall faster: relationship depth suffers, and the data structure that worked at 100 contacts becomes hard to navigate at 500. What vanishes in a Notion export: The data loss map When you export a Notion database to CSV, you get a spreadsheet. But Notion's most powerful feature—linked records and relationship metadata—does not export cleanly. Fields that survive the export (mostly intact) Text fields (name, email, phone) Single-select or multi-select (status, tags) Dates (created, last contact) Numbers (deal amount, interaction count) Checkboxes (opted-in, customer, prospect) Fields that export broken or orphaned Linked records: Export as text IDs or lose the link entirely. A contact linked to 3 deals exports as a comma-separated list of Notion IDs with no metadata about deal status, amount, or timeline. Rollup fields: Do not export. You lose calculated fields like "total deal value for this contact" or "days since last interaction." Relation backreferences: If Company A links to 50 Contacts, those 50 contacts will not know which company they belong to unless you explicitly linked Company → Contact. The reverse link orphans on export. Formulas: Export as the calculated value (good), but lose the logic. If you rebuild in a new CRM, you must rewrite the formula. Interaction history (rich text blocks, nested relations): Export as flattened text or vanish completely. File attachments: Export as URLs. If you move to a new CRM, those URLs often break. In a live test: 500 contacts exported to CSV. 120 of them (24%) had broken or missing links to their primary deal. 87 contacts (17%) had blank company associations because the company-to-contact relationship was one-directional in Notion. Why Notion breaks at 500 but a purpose-built CRM doesn't Notion is a spreadsheet with superpowers. A CRM is built to handle relationship depth, scale contacts without slowdown, and store interaction history as a first-class citizen—not a linked record. When you export from Notion at 500 contacts, you're exporting a flattened structure. A purpose-built CRM platform stores contacts, companies, and deals as interconnected objects with native relationship queries. That means: Searching "all contacts at Company X who touched a deal in the last 30 days" takes milliseconds, not minutes. Relationship metadata (when the link was created, who owns the relationship, deal stage) persists in the export. Interaction history is logged natively, not stitched together from linked records. A contact can belong to multiple deal pipelines without breaking the graph. Exports maintain referential integrity—no orphaned records. At 1,000 contacts, Notion's slowdown becomes a business bottleneck. Your sales team waits for views to load. Your forecasting breaks because rollups fall behind. Your export becomes a recovery nightmare. The 14-day migration plan: Zero orphans, full context Migrating out of Notion means more than moving CSV rows. You're rebuilding relationship depth. This plan assumes you're moving to a CRM that supports rich relationship mapping and native interaction logging. Days 1–2: Audit and map Export all Notion tables (Contacts, Companies, Deals, Interactions) as separate CSVs. In a spreadsheet, map every linked-record relationship: Contact → Company, Contact → Deal, Deal → Owner. Note which relationships are one-directional (broken on export). Identify orphaned re