You've hit 280 contacts in Notion. Search takes 12 seconds. Filters stall. Relationships—the stuff that should link deals to people to companies—start rendering as plain text or break entirely. Then you export to back up or migrate, and nine critical fields vanish. Gone. You're left stitching them back together in a spreadsheet while your sales team sits idle. This isn't a Notion bug. It's a design ceiling. Notion is a database tool that happens to look like a CRM. It was never architected to handle the relational depth, query speed, or data integrity that sales work demands at scale. Once you've built real workflow on it, migrating costs time and carries real risk of data loss. This guide documents exactly where Notion CRM breaks, which fields orphan on export, and a concrete 30-day migration path that recovers your data without losing relationship context. Where Notion CRM hits its performance wall The slowdown is real and measurable: 300+ contacts: Filter queries jump from <1s to 3–8s. Sorting by multiple properties lags. Relation field rendering stalls. 500+ contacts: Database views become unusable. Rollup calculations timeout. Linked database blocks refresh in 15–20 seconds. 1,000+ contacts: The database is functionally broken for daily sales work. The root cause is Notion's underlying architecture. Notion databases are hierarchical document stores, not relational databases. When you add a relation (like "deals for this contact"), Notion must traverse every record in both tables to render that view. At scale, this becomes expensive. Notion's servers then throttle queries to protect their infrastructure, and users experience timeouts. For comparison: a purpose-built CRM like Orin's CRM uses indexed relational queries that return the same data in 50–200ms, even at 10,000+ contacts. The difference is architectural, not just performance tuning. The nine fields that vanish on Notion export When you export a Notion database to CSV, the platform claims it's "lossless." It's not. Nine categories of critical CRM data either corrupt or disappear entirely: 1. Relation fields (deals, companies, contacts) Export behavior: Lists collapse to pipe-delimited text. A contact with five deals shows as Deal-001 | Deal-002 | Deal-003 . Reimporting requires parsing logic, and the back-link (from deal to contact) is lost entirely. You must rebuild the relation in your new system. 2. Rollup fields (count, sum, max date) Export behavior: Values freeze at export time; they don't recalculate on import. If you had a rollup counting "deals in this quarter," that number is now static text. When you add a new deal in your new system, the count won't update. 3. Formula fields (revenue recognition, age, status logic) Export behavior: Formulas export as their computed values, not their logic. The formula itself is lost. Your new system won't auto-calculate, and you can't audit how the value was derived. 4. Button fields (Zapier integrations, webhook triggers) Export behavior: Buttons don't export at all. Any automation or workflow trigger tied to a button disappears. If you had a "Send quote" button that fired a Zapier workflow, that automation is severed. 5. Lookup fields (inherited properties from relations) Export behavior: Lookups export as values, not queries. A lookup that pulled the company size from a related company record exports as a static number. It won't update if the source record changes. 6. Status and select fields with conditional display or color logic Export behavior: Status values export as text; color coding and conditional formatting are lost. Your CRM loses visual workflow signals. A pipeline view organized by deal stage becomes a plain text column. 7. People / User fields (assigned rep, last contact person) Export behavior: User emails export, but permission and team context is lost. If you exported and re-imported into a different Notion workspace, users would be unresolved emails, not actual team members. 8. Rich text and file attachments (notes, contracts, case files) Export behavior: Rich text flattens to plain text; file links break. Any embedded formatting (bold, links, bullet points) in a deal summary note is lost. Attachment links point back to the old Notion instance and will break after domain migration. 9. Timestamp fields (created, last modified) with timezone context Export behavior: Timestamps export in UTC but lose your local timezone context. A field showing "contacted on 2024-03-15 14:30" becomes ambiguous on reimport. If your new system interprets it as UTC and you're in SGT (UTC+8), the time shifts by 8 hours. The upshot: A Notion export is a snapshot of data, not a portable database. Relationships, logic, and automation are stripped away. You're exporting a list, not a system. Why 300 contacts is really the ceiling Beyond slowdown, three other limits compound: Team collaboration breaks: At 300+ records, simultaneous edits cause sync conflicts. Notion's offline-first sync model doesn't hand