You've built a solid Notion CRM. Three years of relationship notes, deal stages, contact hierarchies, timeline history. Then you hit 500 contacts and Notion stops scaling. Export time arrives. The CSV lands in your inbox. You open it. Half your relational data is gone. This is not a Notion bug. It's a design choice: Notion exports flatten structures that were never meant to be flat. Agencies and consultancies hit this wall hardest—they live in relationship graphs. When you migrate to a real CRM like Orin, you need to know exactly which nine fields will orphan, how to recover them, and the 30-day rebuild sequence that gets you whole again. The nine fields Notion forgets to export When you export a Notion database with linked records, rollups, formulas, and relation properties, here's what actually makes it into the CSV: Linked records (many-to-many relations): If you link a contact to a deal, Notion exports the deal ID—nothing else. The actual link vanishes. You get an orphaned reference, not a connection. Rollup fields: If you've calculated "total contract value from linked deals" using a rollup, that value disappears. You get blank cells. The formula logic doesn't export at all. Timeline view sort order: Notion stores deal progression as a timeline. The export gives you raw dates, but the sequencing logic—which determines deal stage flow—vanishes. Relation depth (company → contact → contact → opportunity): Multi-level hierarchies collapse to a single flat list. Your org chart becomes a pile of names. Kanban stage mappings: Your deal stages (Pipeline → Qualified → Proposal → Won) exist as database properties, but the visual workflow logic doesn't translate. You inherit raw text labels, not stage definitions. Filter and sort configurations: Any saved view that filters by "Stage = Proposal AND Contact.Company.Industry = Tech" is gone. You start over with an unsorted list. Custom formula outputs: If you've built formulas that concatenate fields, calculate days-in-stage, or generate deal scores, those calculated values don't export as static data. The formulas themselves are invisible. Audit trail and revision history: Notion stores who changed what and when. None of that metadata makes it into the export. You lose deal movement history. Rich text formatting and embedded media: Notes with formatting, links, and embedded Loom videos become plain text. Context collapses. The real cost of a flat export isn't the missing data—it's the orphaned relationships. A 500-contact CSV with dangling deal references is harder to fix than a smaller, cleaner dataset. Why it happens: the structural mismatch Notion treats a database as a flexible document store. Relations are annotations on documents, not first-class objects. When you export, Notion can only serialize what exists in a single row. A many-to-many relation between a contact and five deals has to pick: link all five in one cell (messy), or export only the first one (lossy). Notion chooses lossy. A proper CRM—including Orin's CRM —stores relationships as edges in a graph. A contact-to-deal link is a persistent object with its own metadata: created date, deal stage, revenue impact. When you export from Orin, that edge travels with you because it's not dependent on the row it came from. Notion's export limitation isn't a bug; it's a consequence of design decisions that prioritize flexibility over relational integrity. If you've been using Notion as a CRM because you liked the flexibility, you were always going to hit this wall. The recovery workflow: three phases Phase 1: Triage and audit (days 1–3) Before you touch anything: Export Notion data to CSV. Do this three times from three different views (All Contacts, All Deals, All Companies) to capture as much as possible. Import into a staging sheet. Use Google Sheets or Excel to create three sheets: Contacts (with all contact-level data), Deals (with all deal-level data), Companies (with all company-level data). Do not combine them yet. Identify orphaned links. Search for cells with deal IDs or company IDs that reference rows outside the current sheet. Flag these with a new column: "Needs Relink." Count formula casualties. For any column that had a Notion formula (days-in-stage, deal score, etc.), spot-check five rows. If values look impossible (negative days, zero scores), mark the column as "Recalculate." Map your Kanban stages. List every deal stage that appeared in Notion (Pipeline, Qualified, Proposal, Won, Lost). These become your Orin pipeline stages. Don't rename them yet; just map them 1:1. Phase 2: Clean and enrich (days 4–14) Now rebuild what the export couldn't capture: Restore company → contact relationships. You likely have a "Company" field on each contact row, but the inverse (contacts linked to company) is gone. Use a VLOOKUP or a pivot table to rebuild this. Create a new sheet: "Company Relationships" with company name, all linked contacts, contact role, and relationship strength (primary, secondary, in