We migrated a sales team from Affinity to test relationship depth, export fidelity, and the hard ceiling where each platform stops working. What we found: relationship graphs don't travel well. One platform exports a flattened contact list. Another preserves networks but bleeds data on growth. One rebuilds what others lose—but the rebuild took 60 days and exposed five critical gaps that the exporter never mentioned. How we tested relationship mapping We took a 180-contact Affinity workspace with 47 active deals, each tagged with decision-maker trees, influencers, champions, and stakeholders. These weren't flat contact-deal links; they were networks: person A influences person B, who reports to person C, who has signed three prior deals with your company. We exported the relationship graph from each platform, re-imported into a staging Orin workspace, and mapped what arrived intact, what got orphaned, and where each platform's native data model forced us to flatten or drop context. The test data Affinity: 180 contacts across 47 deals; relationship tags (decision-maker, champion, influencer, stakeholder); 12 prior interactions per contact average; custom fields for seniority and buying-influence score. PipeDrive: 162 contacts, 41 deals, organization-person linkage native; relationship custom fields mapped to text; prior deal history via activity log. Orin: Fresh import targets with relationship modeling via contact network and deal association ; tested relationship depth and export round-trip. Affinity's export: flat and lossy Affinity exports as CSV. The moment you pull the data, your relationship graph collapses. The export gave us: Contact ID, name, email, phone, company. Custom fields (seniority, buying-influence score)—present. No relationship edges. No "person A influences person B." No decision-maker hierarchy. Deal associations showed only the primary contact per deal; secondary stakeholders orphaned unless manually linked in a comment. Activity timestamps preserved, but context (which contact initiated, who was copied, what relationship role they played) flattened to description text. What we lost: Of 47 deals, 31 had 2+ stakeholders. Affinity's export dropped stakeholder relationships entirely. We reconstructed 18 of them from deal activity threads; 13 stayed orphaned. The buying-influence network—critical for deal strategy—became a scored number on a contact record, not a navigable graph. Affinity exports contacts and deals separately. The relationship map—the actual network that drives deal strategy—does not travel. PipeDrive: linked but limited at scale PipeDrive preserves the person-to-organization link natively. Export includes organization ID on every contact, so stakeholder relationships survive as data—but not as navigable relationships in the re-import. Contact-to-organization linkage preserved in export (foreign key intact). Deal-to-contact associations exported; secondary contacts on deals also exported. Relationship roles (decision-maker, influencer, etc.) stored as custom field text, not first-class entities. Relationship history (who contacted whom, in what order) only in activity feed; not queryable. Where it breaks: PipeDrive's strength is organization-centric deals; it assumes one primary decision-maker per deal and secondary contacts as attributes. At 41 deals with an average 2.8 stakeholders each, the model creaks. Our largest deal had six stakeholders across three organizations; PipeDrive's export forced us to choose a primary contact or lose deal accountability on re-import. We chose the CFO (budget holder) and orphaned two technical influencers. At 162 contacts and $2.4M pipeline, the export was clean. At 400 contacts and $8M pipeline (the team's ceiling), PipeDrive's deal-centric model starts to fragment because relationship depth exceeds the org-primary-contact structure. Orin: relationship depth that scales—if you rebuild first Orin's contact and deal model treats relationships as first-class objects. On import, every contact can have multiple deal associations, and each association stores the relationship role (decision-maker, champion, influencer, stakeholder). The export import preserved structure—but the rebuild was manual. Contacts imported with custom fields intact (seniority, buying-influence score). Deal associations imported with relationship role preserved as metadata. Organization hierarchies imported as contact relationships (org person-to-person links). Activity history imported with context (who acted, when, in what role). The rebuild: On arrival, 31 deals had secondary stakeholders in orphaned state (from Affinity's flat export). Orin's native relationship model could ingest these—but only if we manually re-associated them. We wrote a CSV with contact ID, deal ID, and role, then bulk-imported via Orin's API. Took 60 days because we had to audit each re-association: which stakeholders were real, which were cc'd for noise, which had actually influenced the deal.