You've decided to leave Affinity. Your data export arrives: contacts, interaction history, deal summaries, all there. You import it into your new CRM, tick the completion box, and move on. Three weeks later, your sales team is frustrated. Nobody can see who knows whom. The informal network that drove referrals and deal leverage—the actual relationship graph—is gone. It wasn't lost in a technical failure. It was never exported in the first place. Affinity's relationship graph is its product. It maps how contacts relate to each other, tracks relationship strength through interaction frequency, and surfaces second- and third-degree connections. That logic lives in Affinity's database as metadata layers, interaction patterns, and computed relationship scores. When you export, you get the raw contacts and activity feed. The graph itself—the thing that made Affinity valuable—stays behind. This is not a gotcha or a flaw. It's a design choice. But it means your migration requires a deliberate rebuild phase, not a one-time import. We've migrated a dozen teams out of Affinity into Orin and other platforms. Here's what actually survives the export, what vanishes, and a concrete 60-day script to rebuild your network. What survives the Affinity export The export is not useless. It preserves the foundation: Contact records. Names, email, phone, company, title, and custom fields you added. Fully intact. Interaction history. Timestamps, content, and direction (inbound/outbound) for emails, calls, and notes. This is the raw material of relationship strength. Deal data. Stage, value, close date, deal participants, and custom fields. Account hierarchies sometimes survive, sometimes need manual repair. List membership. Tags or segments you created. These import, though field mappings often need cleanup. If your Affinity export includes the API data (not just the UI export), you may also get relationship IDs and interaction vectors—metadata that hints at the graph, but not the graph itself. Most exports don't include this. Assume they don't. What the export orphans The relationship logic vanishes: Relationship mapping. Affinity computes which contacts relate to each other based on shared companies, interaction patterns, and explicit relationship notes. That computation doesn't export. You have a list of contacts who emailed each other, but no structured "person A introduced person B" or "person C worked with person D at the same company." Network depth. Affinity surfaces secondary and tertiary relationships—the people you can reach through your contacts. That path-finding layer doesn't export. You lose the discovery mechanism that turned a 300-person network into a 3,000-person network. Relationship strength scores. Affinity weights relationships by interaction frequency and recency. You export the raw interactions, but not the computed score or trend. A contact you've emailed 40 times in the last year is indistinguishable from someone you emailed twice, three years ago. Inferred relationships. Affinity infers relationships from shared company IDs, domain overlaps, and event attendance. Those inferences don't export. Your new system sees two people who both worked at Acme, but has no way to know they overlapped. In total, most teams lose 35–50% of the relationship intelligence they relied on. It's not that the data is gone; it's that the logic that made the data navigable is gone. Why a 60-day timeline Rebuilding the graph is not a one-day task. It's not a scripted import. It requires: Days 1–10: Taxonomy definition. How do you want to encode relationships in your new system? Affinity used implicit, computed relationships. Most teams switching to Orin's CRM or similar platforms choose explicit models: relationship types (introduced by, works with, knows socially), strength (strong / medium / weak), and recency flags. Days 11–30: Interaction analysis. Pull your interaction history from the export. Use it as evidence. Who exchanged 20+ emails? Likely strong relationship. Three emails over five years? Weak or dormant. Sort and bucket your top 500–1000 relationships by interaction volume. Days 31–45: Structured entry. Map relationships in your new system using a consistent format. If you choose Orin's native relationship graph , you can encode these as linked contacts with relationship type and strength metadata. Days 46–60: Validation and live switch. Spot-check 10% of relationships. Review with your sales team. Make corrections. Go live with the new system as your source of truth. If you skip this and import raw contacts only, you'll spend the next three months with your sales team manually re-discovering relationships. A deliberate 60-day rebuild frontloads the pain and gives you a usable network on the other side. Building the rebuild script in Orin Here's a concrete workflow: Step 1: Import contacts and interaction history (Days 1–5) Export your Affinity contacts and activity feed (or use the API if available). Import contacts