Affinity shines at relationship intelligence—it maps wealth networks, tracks investor patterns, and captures the connective tissue between your contacts. Then you hit export, and that intelligence evaporates. You get contacts, companies, and notes. What you lose is the graph itself: who knows whom, what leverage each connection carries, and which relationships unlock your next deal. This is not a limitation of Affinity. It's a hard truth about exporting relationship data from any specialized platform. The relationships that live in the interface—the lines between nodes on the map—often don't translate to rows and columns. You export the nodes. The edges stay behind. If you're moving to Orin's CRM and relationship tracking , or if you're auditing whether Affinity's export actually captures what you need, this playbook walks you through what vanishes, how to validate the gap, and a concrete 60-day rebuild that keeps your relationship intelligence intact. What Affinity actually exports—and what stays behind Let's be specific about what leaves Affinity and what doesn't. What exports cleanly Contacts: Name, email, phone, title, company association. Standard contact fields port with minimal friction. Companies: Company name, domain, location, industry tags. Mostly intact, though custom company attributes vary. Deals: Deal name, value, stage, close date, assigned owner. Pipeline structure moves, though probability weighting may not. Notes and activity: Timestamped notes export as text. Task history and email logs export as summaries or not at all, depending on your export format. Custom fields: Any field you manually created in Affinity exports if your export tool supports them. If not, you lose the data entirely. What flattens or vanishes Relationship edges: The fact that Alice knows Bob, and that Bob introduced you to Carol, lives in Affinity's interface but not in the export file itself. You must manually rebuild these connections or infer them from notes. Relationship roles: Affinity lets you tag a contact as 'decision maker,' 'champion,' 'influencer,' or custom types. These tags may export as custom fields, but the relationship context—"Bob is the decision maker on this specific deal "—flattens into a single attribute, losing the deal-specific role. Interaction patterns: Affinity can infer engagement (email opens, call frequency, meeting attendance). Standard exports don't include these signals. You get notes, not analytics. Network affinity scores: Affinity's proprietary scoring of how central a contact is to your network doesn't export. You rebuild from degree count (how many connections) alone. Warm intro logic: If Affinity mapped "Alice → Bob → Carol," that transitive path is inference. It's not stored in a field; it's computed on demand. Your export has three separate contact records with no path information between them. The core truth: Affinity exports a node list. Rebuilding relationship intelligence means reconstructing the edges and weighting them correctly. Audit step one: validate what actually exported Before you rebuild, you need to know what you lost and what you kept. This audit takes 4–6 hours for a 500-contact book and catches silent data rot early. The validation checklist Row count vs. Affinity UI: Export your data to CSV. Count rows. Cross-check against Affinity's contact total. Any discrepancy means filtered contacts or export errors. Document the delta. Email and phone coverage: In the CSV, count non-empty email cells and non-empty phone cells. Compare to your Affinity contact count. Affinity allows many contacts without email (e.g., decision makers you know through intermediaries). If you're under 75% email coverage, you're missing cold-start data for outreach rebuilds. Custom field survival: List every custom field you created in Affinity (e.g., 'relationship_strength', 'investment_thesis', 'warm_intro_source'). Check the CSV header row. For each missing custom field, document what it contained. This is your rebuild checklist. Deal-contact associations: For your top 20 deals, cross-check the CSV deal export. For each deal, verify that every contact in Affinity's visual "people" list for that deal appears in the CSV. Note any contact IDs missing from the deal association table. These are orphaned relationships. Note completeness: Sample 10 contacts from the CSV. For each, read the notes field. Compare to the original contact record in Affinity (if you still have access). Are all notes present? Are timestamps intact? Do notes reference other contacts by name but not by ID? (This matters for rebuilding relationships later.) Activity timestamps: If you exported activity logs, pick 3 contacts and verify the last activity date in the CSV matches Affinity's UI. Date mismatches reveal export window issues. The 60-day rebuild playbook in Orin Now you import into Orin's CRM and rebuild relationship intelligence systematically. The 60 days breaks into four phases. Days 1–15: Import and stabilize