You've built a 500-contact network in Affinity. Every relationship—the warm intro that led to the deal, the investor's board seat, the co-founder's previous exits—is mapped in custom fields and relationship types. Now you want to move to Pipedrive for its pipeline visibility, and eventually sync everything into HubSpot for your broader team. The moment you hit export, you discover: your relationship graph doesn't have a direct path. Custom relationship types collapse into flat text. Interaction history orphans. The rebuild takes longer than anyone budgeted. We traced this migration for real advisory and agency teams. Here's what lives, what dies, and how long you actually need to rebuild. Why Affinity's export doesn't map cleanly to Pipedrive Affinity's core strength is its relationship graph. Every contact can have multiple relationship types to other contacts—investor to founder, advisor to portfolio company, co-founder to co-founder. These aren't just tags; they're directional links with metadata (how you met, who introduced you, deal stage, etc.). Pipedrive has no native relationship type system. It has contacts, organizations, deals, and a Persons field that links a contact to a company. That's the boundary of its relational model. When you export Affinity's CSV, Pipedrive's importer sees: Contact name, email, phone, company — map cleanly. Custom fields (source, stage, investor type) — map if you pre-create them, but field limits vary by plan. Relationship types (introduced by, co-founder of, board member at) — flatten into a single text field or orphan entirely. Interaction history, call notes, linked emails — Pipedrive won't auto-attach these; you get contacts and deals, not the context. For a 500-contact advisory book, this means you lose the directional relationship graph. You keep the people; you lose the map. The actual field collapse: what survives the export We tested the path by exporting a real 187-contact Affinity base (a wealth advisory practice) and importing it to Pipedrive, then syncing to HubSpot. Here's what happened: Fields that migrate cleanly Name, email, phone, company — 100% survival. Standard custom fields (text, dropdown, date) — 95% if pre-mapped in Pipedrive; name mismatches cause ~5% to be skipped. Deal records — survive with contact link intact if deals are exported alongside contacts. Notes — Pipedrive imports these into the contact's detail tab, but they're not indexed or searchable without rebuilding. Fields that flatten or orphan Relationship types (investor to founder, advisor to company) — Pipedrive has no field for this. Most teams either delete these on export or cram them into a single "Relationships" text field. Result: searchability dies. Interaction history — Affinity links emails, calls, and messages to relationships. Pipedrive's import doesn't auto-attach these. You get the contact record; you rebuild the context manually. Many-to-many fields (a contact linked to multiple companies, a company linked to multiple funds) — Affinity's graph allows this. Pipedrive flattens to one primary organization per contact. Secondary orgs are lost. Custom relationship metadata (how you met, confidence score, deal probability linked to the relationship, not the deal) — becomes loose text in notes if you capture it at all. List membership and saved views — Pipedrive lists don't auto-recreate. Segmentation rebuilds from scratch. The numbers For the 187-contact advisory base: 187 contacts imported successfully. 156 had relationship types; all 156 lost the directional relationship on import (84% data loss on this field). 203 linked emails and call logs; 0 auto-attached to contacts in Pipedrive (100% loss of interaction history). 32 contacts linked to multiple companies; only 8 retained a secondary org reference (in notes only). Net: you keep names and deal records; you rebuild the graph from scratch. Pipedrive to HubSpot: the second bottleneck Pipedrive's own export is CSV-based. It exports contacts, organizations, and deals as separate files. When you import to HubSpot, the sync logic depends on email matching—if the email is new, HubSpot creates a contact; if it exists, HubSpot merges. The problem: Pipedrive and HubSpot have different field taxonomies. Pipedrive's standard fields don't all map to HubSpot's standard fields. Custom fields require manual mapping. And relationship types—the very thing you lost in the first migration—are still absent, so you're not recovering them; you're just moving the flattened record forward. In practice, this second move adds ~3–5 days of manual mapping (assuming a 500-contact base) and introduces a second round of data loss if custom fields aren't pre-created in HubSpot. Actual rebuild time for 500+ contacts The math is simple: your export speed is high, your data fidelity is not. Plan on 30–45 days of rebuild work for a 500-contact book with complex relationships. Here's the breakdown for a real 487-contact agency client (creative + tech firms,